EP0667954A4

EP0667954A4 - System and method for testing the integrity of porous elements.

Info

Publication number: EP0667954A4
Application number: EP94900561A
Authority: EP
Inventors: Scott D Hopkins; Daniel W Spencer; Charles P Lipari; George A Altemose
Original assignee: Pall Corp
Current assignee: Pall Corp
Priority date: 1992-11-06
Filing date: 1993-11-08
Publication date: 1995-12-27
Also published as: WO1994011721A1; JPH08503545A; CA2148807A1; EP0667954A1

Abstract

An ultrasonic detection apparatus (1) is disclosed for detecting defects in filters (3). The apparatus (1) operates on a wetted filter (3) and includes a housing (2). The wetted filter (3) divides the housing (2) into an inlet side (7) and an outlet side (8). Both the inlet side (7) and the outlet side (8) may be filled with a gas as with a standard forward flow test apparatus. A microphone (4) is disposed in the vicinity of the wetted filter (3). The microphone (4) receives acoustic signals generated within the chamber (any combination of inlet side (7), outlet side (8), inlet tube (5), or outlet tube (6)) as a result of the increased pressure on the inlet side (7). A signal processing device (9) is also included for analyzing the acoustic signals received by the microphone (4) for determining whether a filter (3) is defective. The invention includes multiple methods for determining whether a filter (3) is defective. One method may include the steps of placing a wetted filter (3) in a test housing (2) to divide the test housing (2) into an inlet side (7) and an outlet side (8), pressurizing the inlet side (7) with gas, measuring both the gas flow and the sound volume on the outlet side (8), and thereby determine whether the filter (3) is defective.

Description

SYSTEM AND METHOD FOR TESTING THE INTEGRITY OF POROUS ELEMENTS

Technical Field of the Invention

The present invention relates to systems and methods which permit the integrity of a porous element, such as a filtration element, to be quickly and reliably tested. In particular, the present invention relates to testing porous elements by detecting sound, for example ultrasound, produced when a wetted porous element is exposed to a differential pressure.

Background of the Invention

In many fluid processing systems involving filtration, there is a requirement to achieve the highest possible assurance of filter integrity and removal efficiency. Examples of such applications include sterilization of parenterals, biological liquids, and fluids used in fermentation processes. Conventional methods utilized to verify the integrity of porous filter media include: microbiological challenge tests, effluent cleanliness tests, particle challenge tests, forward flow tests (including pressure decay tests) , and reverse bubble point tests. The microbiological challenge, particle challenge and effluent cleanliness tests are destructive tests, and cannot be performed in a production environment. The industry-accepted non-destructive tests used to verify the integrity of porous elements are the forward flow test or the reverse bubble point test. Both of these tests are performed by applying a predetermined gas pressure to a wetted filter.

Reverse bubble point testing has been utilized since the 1950s to determine the size and location of the largest pore in a filter element. See, for example, D.B. Pall (U.S. Patent 3,007,334, filed November 30, 1956). A reverse bubble point test (sometimes referred to as a first bubble test) detects defects by looking for bubbles while a wetted filter is submersed in a liquid. With the filter wetted and liquid covering one side of the filter, a gas at constant and/or variable pressure is directed against the other side of the filter.

At a certain pressure, the gas is just able to force the liquid from some of the largest pores in a filter having a homogenous pore structure, and the gas forms bubbles in the liquid covering the filter. This pressure is known as the bubble point of the filter. Of course, as the pressure increases above the bubble point, more and more of the liquid is forced from increasingly smaller pores of the filter and the flow of gas through the filter increases. The bubble point of the filter depends on many factors, but pore size is a dominant factor. Filters having larger pores have a bubble point that occurs at lower pressures when using the same type of wetting solution. If the pressure of the gas is below the bubble point of the filter, very little gas passes through the filter. However, if there are defects in the filter, the gas will pass through the defects in the filter and bubbles will form in the liquid.

Bubbles rising in the liquid can be detected either visually or electronically using a passive sonar device which monitors a sudden increase in the sound intensity caused by bubbles rising and/or collapsing in the liquid. A device which ultrasonically detects a sudden increase in the sound intensity at the bubble point as bubbles rise/collapse in the liquid is, for example, shown by Reichelt, U.S. Patent No. 4,744,240. Reichelt is utilized for determining the bubble point pressure of the largest pores in a homogeneous, non-defective filter. Reichelt is directed to determining the pressure at which a sudden increase in sound volume resulting from an applied pressure at a level in which gas is forced through a plurality of pores in a in-tact filter element having a relatively homogenous pore structure. The apparatus disclosed in Reichelt does not detect defective filters having pin-hole defects. Another problem with the apparatus disclosed in Reichelt is that the liquid media in which the ultrasonic transducer is disposed couples the microphone to sounds down stream of the filter element and throughout the fluid flow system. In this configuration, noise produced by bubbles from, for example, pin-holes becomes lost in the ambient noise of the system and sources of noise outside of the system. Additionally, the Reichelt device only detects a sudden rise in the noise level produced at the bubble point. It was found that many defective filters cannot be detected simply by measuring the sudden rise in the noise level. Reverse bubble point testing has a number of limitations. For filters having a cylindrical configuration, the filter must be rotated as it is observed for the formations of bubbles. The observation of bubbles is hindered by the fact that bubbles may be trapped by the filter as it is placed in the liquid, particularly where the filter has closely spaced pleats. Additionally, depending on the geometry of the filter, type of wetting solution, and the applied pressure, diffusional flow may produce several bubbles per second. Trapped bubbles and the bubbles from diffusional flow may provide a false indication that the filter is defective.

Reverse bubble point testing is not well suited for testing a high volume of filters because the test takes a significant amount of time to complete and is subject to observer limitations. Further, it is exceedingly difficult and of limited value to simultaneously conduct a bubble test and a forward flow test. Additionally, it is impractical to reverse bubble point test a filter in two different directions because of the limitations of the test apparatus and the filter construction (e.g., cartridge filters). Reverse bubble point testing of a filter in an operational environment (on-line testing) is extremely difficult, and impractical under most circumstances. In addition to the above disadvantages, reverse bubble point testing does not provide a quantitative assessment of the filter.

As a result of the limitations of reverse bubble point tests, in the early 1970"s Pall Corporation developed a filtration test known as the forward flow test. See, for example, Dr. D.B. Pall, 1973, "Quality Control of Absolute Bacteria Removal Filters", Parenteral Drug Association, November 2, 1973. Conventional forward flow tests detect defects in a filter by measuring gas flow through a wetted filter. The forward flow test quantitatively measures the sum of diffusive flow and flow through any pores larger than a predetermined size. In forward flow testing, the filter is typically placed in a test housing. The filter is wetted by immersing the filter in a liquid, such as water or alcohol, until all of the pores of the filter are filled with the liquid. The filter may be wetted by, for example, directing deionized water through the filter for a predetermined period of time. A gas is then directed under pressure against one side of the filter and gas flow through the wetted filter is measured by a flow meter, such as a mass flow meter. If the filter has no defects, the gas at low pressures is unable to force the liquid from the pores of the filter, so there is very little gas flow, and typically only diffusive gas flow, through the filter. The wetted filter medium behaves like a sheet of wetting solution whose thickness is equal to that of the filter medium. The gas dissolves in the wetting solution, diffuses through it, and then is released downstream of the filter. At lower pressures, the flow per unit of applied pressure remains substantially constant. The flow measured at the lower pressures can be calculated for a given liquid from the known diffusion constant of the applied gas through the liquid.

At a certain higher pressure, known as the bubble point, the gas is just able to force the liquid from some of the largest pores in the filter, and a sudden increase in the flow of gas through the filter can be detected. Of course, as the pressure increases above the bubble point, more and more of the liquid is forced from the pores of the filter and the flow of gas through the filter increases. The slope of the curve after reaching the bubble point provides a measure of the uniformity of the pore sizes in the filter element. A more accurate measure of the "bubble point", is the quantity "K_L", coined form the term Knee Location, used to indicate the pressure at which the mass flow curve in a forward flow test bends.

Although the forward flow test is extensively used and is very reliable, it nonetheless has certain drawbacks. For example, the test takes a significant amount of time to complete. A large amount of time is required for gas flow to stabilize before testing can even begin. Once the test does begin, it must be conducted over an extended period of time in order to accurately measure the very small flow rates associated with modern filtration devices. Additionally, there may be a loss of accuracy for on¬ line testing using forward flow when several dozen filters are tested in parallel without isolating the individual flows through each of the parallel connected filter elements.

Summary of the Invention

A principal object of the present invention is to alleviate the above-mentioned disadvantages and provide a reliable, economical, and easy-to-operate system and method for testing porous elements. Another principal object of the present invention is to have a relatively short test time and increased accuracy.

Other objects of the present invention include detecting the sound of gas passing through a wetted porous element and analyzing the sound of the gas to discriminate between a defective porous element and a porous element without defects; providing a pass/fail indication for a porous element based on predetermined characteristics of the porous element; discriminating between acoustic signals which indicate that a porous element is defective and acoustic signals which do not indicate that a porous element is defective; minimizing acoustic noise from sources external to the porous element testing system and minimizing electrical noise from circuits contained within the porous element testing system; providing a porous element testing system or method which is compatible with conventional forward flow testing techniques; and detecting defective porous elements regardless of the direction that the porous element is pressurized. Accordingly, the present invention provides a sonic bubble point test conducted by detecting sound, for example ultrasound, produced when a wetted porous element is exposed to a differential pressure that is less than the bubble point (i.e. a pressure less than the pressure at which the liquid if forced from the largest pores of a homogeneous pore structure) . In a preferred embodiment, the sound may be detected while a . fluid in a gaseous phase is disposed over both upstream and downstream surfaces of the wetted porous element. Methods and apparatuses are provided for analyzing the sound detected to provide a quantitative measure of the integrity of the porous element. In a preferred embodiment, sonic signals (preferably air¬ borne sonic signals) from known non-defective filter elements are compared with sonic signals from known defective filter elements to determine test parameters by which a testing apparatus can discriminate between non-defective and defective filter elements. Defects can be detected at differential pressures substantially below the bubble point so that the methods and apparatuses according to the present invention can be conducted simultaneously with forward flow tests. The sonic bubble point test is not subject to the traditional false positive failures attributable to conventional bubble point tests, and is thus capable of being automated.

The present invention provides a porous element testing system for testing a porous element wetted with a wetting solution, the porous element testing system comprising: a housing having a first side and a second side, wherein the first side is divided from the second side by the wetted porous element and wherein the first side and the second side are both filled with a gas; a differential pressure generator for generating a differential pressure across the wetted porous element; a transducer disposed in the vicinity of the porous element for receiving acoustic signals generated within the housing; a signal processing device, coupled to the transducer, for analyzing the acoustic signals received from the transducer and for determining whether the porous element is defective.

A method for determining whether a porous element is defective comprising: wetting a porous element with a wetting solution; generating a differential pressure between a gas on a first side of the wetted porous element and a gas on a second side of the wetted porous element; monitoring sound levels emanating from the vicinity of the porous element; and determining, as a result of the sound levels, whether the porous element is defective.

A system for testing a wetted porous element, the testing system comprising: a housing having first and second sides, wherein the first side is dividable from the second side by the wetted porous element; differential pressure generator for applying a differential pressure between the first and second sides of the housing; a sound transducer for receiving acoustic signals generated within the housing; and a gas flow meter arranged to monitor gas flow between the first and second sides of the housing. A method for determining whether a porous element is defective, comprising: wetting a porous element with a wetting solution; creating a differential pressure across the wetted porous element; monitoring acoustic signals generated in the vicinity of the wetted porous element; and measuring gas flow through the wetted porous element. A porous element testing system for testing a wetted porous element, the porous element testing system comprising: a differential pressure generator arranged to generate a differential pressure less than the bubble point pressure across a wetted porous element; a transducer disposed in the vicinity of the wetted porous element to receive acoustic signals; a signal processing device, coupled to the transducer, for analyzing the acoustic signals received from the transducer and for determining whether the wetted porous element is defective.

A method for determining whether a porous element is defective comprising: wetting a porous element with a wetting solution; exposing first and second sides of the porous element to a differential pressure less than a bubble point pressure; monitoring sound levels adjacent to the porous element; determining, as a result of the sound levels, whether the porous element is defective.

Apparatus for testing a wetted porous element comprising: a transducer positionable in the vicinity of the wetted porous element to receive acoustic signals and a signal processing device coupled to the transducer to count pulses in the acoustic signal and determine whether the porous element is defective in accordance with the pulse count.

A method for testing a wetted porous element comprising counting acoustic pulses emanating from the wetted porous element and determining whether the porous element is defective in accordance with the pulse count.

Brief Description of the Drawings

Figure 1 is a plan/block diagram of the porous element testing system according to one embodiment of the invention.

Figure 2 is a sectional view of one embodiment of the test housing shown in Figure 1.

Figure 3 is a partial block/partial schematic diagram of one embodiment of the signal processing device shown in Figure 1.

Figure 4A and 4B are a schematic diagrams of embodiments of the conditioning circuit of Figure 3.

Figure 5 is a schematic diagram of an embodiment of the pulse width detector of Figure 3. Figure 6 is a schematic diagram of one embodiment of the output logic of Figure 3.

Figures 7A-G are graphical representations of a testing cycle according to the first embodiment of the signal processing device illustrated in Figure 3. Figures 8A-G depict simplified graphical representations of the responses of the apparatus of Figure 1 to various filter conditions.

Figure 9 is a graphical representation of an alternative testing cycle of the apparatus illustrated in Figure 1.

Figure 10 is a partial block/partial schematic diagram of one embodiment of the pulse count circuitry shown in Figure 3. Figure 11 is a graphical representation of a test cycle according to one configuration of the pulse count circuitry shown in Figure 10.

Figure 12 is a block diagram of an alternative embodiment of the testing system shown in Figure 1. Figure 13 is a graphical representation of a testing cycle of a non-defective filter according to the embodiment of the testing apparatus illustrated in Figure 12.

Figure 14 is a graphical representation of a testing cycle of a defective filter according to the embodiment of the testing apparatus illustrated in Figure 12.

Figure 15 is a graphical representation of a typical front panel screen displayed during a testing cycle of a filter according to the embodiment of the testing apparatus illustrated in Figure 12.

Figure 16 is a graphical representation a typical software program employed by the embodiment of the testing apparatus illustrated in Figure 12.

Description of Embodiments

Systems and methods embodying the present invention may be used to quickly and reliably test the integrity of a wide variety of porous elements. For example, the porous element may comprise a porous medium such as a porous membrane, a porous fibrous sheet or mass, porous, hollow fibers, a woven or non- woven mesh, and/or a porous sintered or non-sintered structure. The porous element may also comprise a cartridge or module having one or more of the following components: a porous medium, a porous support, drainage material, a support plate, an end cap, a core, and a cage. Further, the porous element may comprise an assembly including, for example, a housing containing a porous medium and one or more conduits or fittings associated with the housing. The porous element may have any desired geometry; for example, it may be configured as a solid or hollow cylinder, a disk, or a flat or non-flat sheet. In addition, the porous element may have any desired pore size and distribution; for example, it may be microporous or ultraporous and it may have a uniform or graded pore distribution.

Systems and methods embodying the present invention may be used to determine the presence of a wide variety of defects in porous elements. These defects include not only pin holes or tears in a porous medium but also irregularities in the porous medium such as uncommonly large pores. These defects also include faulty bonds, for example, between a porous medium and an end cap, cracks or holes, for example, in an end cap or a housing, and other flaws through which gas may pass.

As shown in Figures 1 and 2, one example of a porous element testing system 1 embodying the invention incorporates a housing 2 coupled to an inlet tube 5 and an outlet tube 6. The housing may be the housing normally containing the porous element during routine filtration operations. Alternatively, the housing may be a housing used solely for testing and having a geometry suitably adapted to the geometry of the particular type of porous element to be tested. For example, the housing may simply comprise two impervious plate-shaped pieces between which a sheet of a porous element is sandwiched. Using this arrangement, bulk filtration material can be tested during the manufacturing process. In the illustrated porous element testing system 1, the housing 2 has a generally cylindrical geometry and is adapted for testing a hollow, cylindrical filter 3 having a porous medium 11. The outlet tube 6 may incorporate a fitting 14 that has a first tube which is connected to the outlet tube 6, a second tube which may be coupled to a microphone 4, a third tube which provides a drain 13, and an optional fourth tube coupled to a flow meter 55. The fitting 14 may include an elastomeric fitting 43 in which the microphone 4 is placed, and serve to acoustically isolate the microphone 4 from external acoustic signals. The fitting 14 may also seal the microphone to the outlet tube to prevent gas bypass. The microphone 4 is preferably coupled to a signal processing device 9.

The filter 3 is preferably first wetted and placed in the housing 2. The filter 3 is positioned such that it separates the housing 2 into an inlet side 7 and an outlet side 8. A gas control arrangement 54 maybe coupled to an inlet tube 5. The inlet tube 5 is, in turn, coupled to the inlet side 7 of the housing 2. The outlet tube 6 is coupled to the outlet side 8 of the housing 2. Although the illustrated microphone 4 is coupled to the fitting 14 , it may alternatively or additionally be placed in number of several locations. For example, the microphone may be placed within the outlet tube 6, preferably near the junction of the outlet tube 6 and the outlet side 8 of the housing 2, in the outlet side 8 of the housing 2, in the inlet side 7 of the housing 2, and/or in the inlet tube 5. In many embodiments it is preferable to locate the microphone 4 within a line-of-sight of the porous element. Placing the microphone 4 within a line-of-sight of the porous element reduces distortions and increases the sound pressure level at the microphone. In the most preferred embodiment, the microphone is located in the inlet tube or the outlet tube to provide improved discrimination between defective and non-defective porous elements.

The signal processing device 9 analyzes the sound pressure level detected by the microphone in order to perform a variety of desirable functions. For example, the signal processing device 9 most preferably discriminates between defective porous elements and porous elements without defects. It also may be used to determine characteristics of the porous element, such as pore size, or characteristics of the defect, such as size, or other anomalies such as improper wetting.

Although the signal processing device 9 may be configured in a variety of ways, one example of the signal processing device 9 is shown in Figure 3. The microphone 4 detects sound pressure levels within the housing 2 and outputs a signal indicative of the sound pressure level detected. The signal from the microphone 4 is amplified by a pre-amplifier 15 and output as a pre-amplified signal. An adjustable bandpass filter 16 operates to condition the pre- amplified signal. In one embodiment, the bandpass filter 16 filters the pre-amplified signal so that a narrow frequency band is output as a filtered signal. A variable gain amplifier 17 may be provided to amplify the filtered signal to output an amplified signal. The amplified signal is received by a conditioning circuit 18 and by pulse count circuitry 60. An example of one embodiment of the pulse count circuitry 60 is shown in Figure 10 and discussed in more detail below. The conditioning circuit 18 reshapes the amplified signal and outputs a conditioned signal to an output driver 20 and threshold comparator 21. The output driver 20 couples the conditioned signal to a visual display device such as a chart recorder 40. The chart recorder 40 may enable an operator to visually detect defects. However, operator analysis of charts produced by the chart recorder require a substantial amount of time and sophistication on the part of the operator and, therefore, the chart recorder is a less preferred embodiment. Additionally, in many circumstances, an operator may not be able to distinguish a faulty filter from an operational filter because of noise without the assistance of additional processing of the signal. A chart recorder may not be sufficiently precise to display low amplitude or short duration responses caused by small defects; however, the chart recorder is certainly^' adequate to display a response associated with the bubble point of the porous element.

The conditioned signal is input into a first input of the threshold comparator 21. A user-variable threshold voltage is input into a second input of the threshold comparator 21. The threshold comparator 21 is coupled to a variable pulse width detector 46 which detects when a signal from the threshold comparator 21 is activated for a predetermined period of time. An example of one embodiment of a pulse width detector 46 is shown in Figure 5. The variable pulse width detector 46 may output a pulse width exceeded signal to a threshold indicator 22 and output logic 27. The output logic 27 may, for example, be constructed as shown in Figure 6. Output logic 27 may be coupled to a fail indicator 23, a pass indicator 24, a stabilization indicator 25, and a testing indicator 26.

Any suitable control circuitry for the signal processing device 9 may be provided. In the embodiment illustrated in Figure 3, a 115 volt power source 39 supplies power to a power supply 38 via a voltage transformer 42. The power supply 38 may provide operational voltages to the components in the signal processing device 9. A 60 Hz input signal from the voltage transformer 42 may be coupled to a clock generator circuit 35. The clock generator may, for example, input the 60 Hz (or 50 Hz) input signal, divides the input signal by 60 (or 50) , and produce a 1 Hz clock signal. The 1 Hz clock signal may be input into a master counter 34. Alternatively, the master counter may receive a constant or variable clock having any suitable frequency from a standard oscillator circuit. A start input may be coupled to start logic 36 via an optical isolator 37. In the illustrated embodiment, the start logic 36 is coupled to the master counter 34 via a reset signal and coupled to the clock generator 35 via a sync signal. The master counter 34 may be coupled to a display driver 33, a total test time comparator 29, and to a stabilization time comparator 30 via a current count bus signal. The display driver 33 is coupled to a display 32. The total test time comparator 29 has a first input connected to the current count bus, a second input connected to a total test time set switch 28, and an output connected to the output logic 27 and the master counter 34. The stabilization time comparator 30 has a first input connected to the current count bus, a second input connected to a stabilization time set switch 31, and an output connected to the output logic 27.

In a preferred mode of operation, the porous element, e.g., the filter element 3, is wetted with a suitable wetting solution such as water and/or alcohol. A preferred wetting solution for many hydrophobic filters is a mixture of water (75 parts by volume) with tertiary butyl alcohol (25 parts by volume), known by the trade designation Pallsol. The filter 3 may be wetted before or after it is placed in the housing 2. With the wetted filter 3 sealed within the housing 2, a gas is then introduced into the housing 2 through either the inlet tube 5 or the outlet tube 6. For many tests, the gas is preferably introduced through the inlet tube 5. Alternatively, the filter may be pressurized from the inside out by introducing the gas into the housing 2 through the outlet tube 6.

In one embodiment, the gas is introduced into the housing 2 by the gas control arrangement 54, which may be coupled to or independent of the signal processing device 9. In a preferred embodiment, the gas control arrangement may be part of a forward flow test system and the porous element testing system may be coupled to and function in conjunction with the forward flow test system. In any event, the housing 2 may be pressurized gradually, e.g., by ramping the pressure, or, more preferably, it may be stepped to a predetermined value. The predetermined value may range from 50% to 95%, or preferably from 60% to 90% or more preferably from 75% to 85%, or most preferably the predetermined value is 80% of the predicted bubble point of the filter 3.

At pressures below the bubble point, only a small amount of gas should diffuse through the porous medium 11 of the filter 3 and the microphone 4 should not detect the sound of the gas passing through either the pores of the porous medium 11 or a defect in the filter 3. If the sound of gas passing through the pores or a defect is detected by the microphone 4 at pressures below the bubble point, then the filter 3 is likely to be defective. For example, the porous medium itself may be defective due, for example, to abnormally large pores, a hole, or a tear. Alternatively, the filter 3 may be defective due, for example, to a defective bond between the porous medium and the end caps or due to a crack in the end caps.

In the embodiment shown in Figure 3, the start logic 36, in response to the start input, operates to reset the master counter 34, and subsequently enables the clock generator 35 to supply the 1 Hz clock to the master counter 34. The start input may be provided manually by the user or occur automatically as the housing 2 is pressurized. For example, the start input may be received from a 115 volt power source which is energized once the gas control arrangement 54 pressurizes the housing 2. The master counter 34 may count up in one second intervals, or at a lessor or greater interval. In the preferred embodiment, the current count bus signal output by the master counter

34 is indicative of the number of seconds that have been counted since the test was initiated by the start logic 36. The display driver 33 receives the current count bus signal indicative of the number of seconds elapsed, and displays the number of seconds elapsed on the display 32. The display 32 may provide a visual indication of the progress of the ultrasonic test.

The stabilization comparator 30 receives the current count bus signal from the master counter 34 and compares the current count against a stabilization time indicated by the stabilization time set switch 31. The time indicated by the stabilization time set switch 31 may vary depending, for example, on characteristics of the porous element such as pore size, the size and physical configuration of the porous element, etc. Exemplary stabilization times may be in the range up to about 15-20 seconds. When the stabilization comparator 30 detects that the predetermined stabilization time has been reached, the stabilization comparator 30 outputs a stabilization time complete signal to the output logic 27. The stabilization complete signal causes the output logic

27 to begin monitoring the pulse width exceeded signal for indicating a failure whenever this signal is activated after the stabilization period.

The total test time comparator 29 receives the current count bus signal from the master counter 34 and compares the current count against a total test time indicated by the total test time set switch 28. The time indicated by the total test time set switch

28 may also vary depending, for example, on the characteristics of the porous medium and the desired parameters of the test. Exemplary total test times may be in the range up to about 45-50 seconds. When the total test time comparator 29 detects that the predetermined total test time has been reached, the total test time comparator outputs a total time complete signal to the output logic 27 and the master counter 34. The master counter 34 may then inhibit further counting in response to receiving the total test time complete signal so that counting will not resume until the master counter 34 has been again reset by the start logic 36.

Throughout this timing sequence, the microphone 4 outputs a signal responsive to the sound detected by the microphone 4. The microphone 4 may be any suitable transducer for converting sound pressure into electrical energy. If the microphone 4 is of the piezo-ceramic type, then the microphone 4 will only be resistive at the resonant frequency and at an anti- resonance frequency. To optimize electrical to mechanical efficiency for transmitting, a piezo- ceramic transducer is preferably be operated at the resonance frequency. To optimize mechanical to electrical efficiency for receiving, the piezo¬ electric transducer is preferably operated at the anti-resonance frequency.

The microphone 4 is preferably a piezo-electric crystal transducer and the optimum mechanical to electrical frequency of the microphone 4 is preferably in the range of about 30 to about 50 KHz and more preferably about 40 KHz. A frequency above 30 KHz avoids ambient acoustic noise and a frequency below 50 KHz avoids inherent electrical noise generated by high frequency circuity. Under normal test conditions, the signal from the microphone is on the order of 1 microvolt (1 uV) at a frequency of about 40 KHz.

It has been found that drops of liquid on the microphone 4 may reduce the sensitivity of the microphone 4. Therefore, it is desirable to shield the microphone 4 from the wetting solution or other liquids by, for example, placing the microphone 4 downstream from the drain 13, heating the microphone 4 by means of a heating element (not shown) in order to evaporate the liquid, polishing the microphone dome to a highly smoothed surface, and/or by applying a voltage to the microphone to vaporize any liquid on the microphone by sonic vibrations.

The microphone may be constructed with a highly smoothed or polished surface obtained, for example, by electro-polishing the dome of an ultrasonic transducer. The polished surface prevents liquids from being trapped in pores or depressions on the surface and results in a hydrophobic microphone. By hydrophobic it is meant that any wetting fluid which comes in contact with the microphone beads and rolls off without wetting the surface of the microphone. The highly polished surface preferably has no more than 10 micro inches of discontinuity (Ra = 10) , or more preferably no more than 8 micro inches of discontinuity (Ra = 8) , or even more preferably no more than 6 micro inches of discontinuity (Ra = 6) , or most preferably no more than 4 micro inches of discontinuity (Ra = 4) . In one embodiment, electro- polishing of the microphone is continued until the microphone 4 is optimized to maximize the mechanical to electrical efficiency at a predetermined frequency of, for example, at about 40 KHz. In a preferred embodiment, the dome of the microphone is manufactured using stainless steel. A stainless steel microphone may be used in on-line testing applications without danger of contamination or leaching into the system being tested. The microphone 4 is also safely used in a gaseous environment containing volatile solvents and/or wetting solutions since the microphone 4 contains no stored energy.

The pre-amplifier 15 amplifies the signal received from the microphone to a voltage preferably on the order of 1 V. The signal-to-noise ratio (S/N) of pre-amplifier 15 is limited by the self-generated noise of the pre-amplifier itself. The noise is determined by the source resistance, the type of active circuit, and the bandwidth of the signal. It is desirable to maximize the signal-to-noise ratio for the pre-amplifier 15 to increase the sensitivity of the system.

The preferred pre-amplifier 15 may be an operational amplifier commercially available from Precision Monolithics Incorporated (PMI) , Santa Clara, California, or Motorola, under the trade designation OP-27. The OP-27 operational amplifier is an ultra- low noise operational amplifier. In a preferred embodiment, a preamplifier with a S/N of 5.0 or more is preferred for use in pre-amplifier 15. The adjustable bandpass filter 16 is preferably designed to have a center frequency corresponding to the optimum mechanical to electrical conversion frequency of the microphone 4. In a preferred embodiment, the bandpass filter has a bandwidth of 2 KHz and a Q of approximately 20, and is implemented using a biquadratic filter, also known as a state-variable bandpass filter. This filter was found to provide adequate performance with minimum costs, and therefore is a preferred embodiment. It was also found to provide a high probability of detecting a defective porous element, while at the same time minimizing the probability that a properly functioning porous element would be detected as defective (i.e., a false positive result) . A filter with narrower bandwidth (higher Q) could be implemented with increased cost and complexity, resulting in improved S/N. Conversely, a simpler filter or no filter could be utilized, but this may result in an increased S/N and is therefore not preferred for time domain analysis of the signal from the microphone 4. The adjustable bandpass filter 16 has an adjustable center frequency and bandwidth. It has been found that a center frequency of over 35 Khz is preferable because there is much less ambient noise generated by outside environmental factors in this frequency range.

It has been found that the reliability of the system can be greatly increased by reducing the internally generated noise of the pre-amplifier 15 and narrowing the bandwidth of the adjustable bandpass filter 16 around a center frequency corresponding to a frequency at which the porous element generates the largest signal. The frequency at which the porous element generates the largest signal may be the frequency of the sound generated by the largest pores in the porous element. It was found that for some porous elements, the frequency having the maximum amplitude occurs at about 40 KHz.

The variable gain amplifier 17 has a gain that can be varied in response to the position of the microphone relative to the porous element or in response to the characteristics of porous element being tested. The variable gain amplifier 17 allows a plurality of porous element testing systems to be calibrated so that a "single set of test criteria is valid for each of the porous element testing systems.

The conditioning circuit 18 may be constructed to discriminate between noise spikes (having a low voltage and/or infrequent occurrence) and a signal produced by a truly defective porous element. The characteristics of the conditioning circuit 18 may be set such that an isolated short noise spike will not appreciably alter the signal output from the conditioning circuit 18. However, if a plurality of short noise spikes are received within a relatively short period of time, the voltage of the conditioned signal may be conditioned to track the average voltage (baseline) of the noise spikes. In one embodiment, the conditioning circuit has a rise time of 0.5 milliseconds and a fall time of 2.5 milliseconds.

Examples of suitable conditioning circuits are shown in Figures 4A and 4B. Figure 4A shows a half- wave rectifier cascaded with a low pass filter (R9, C3) forming what is commonly known as an average detector.

Figure 4B shows a conditioning circuit including a plurality of resistors in the feed-back circuitry. If resistor R6 and R7 are very small, and resistor R8 is very large, then the circuit in Figure 4B acts as a classic peak detector where the output is a constant voltage corresponding in value to the highest voltage spike (the peak) detected at the input. The classic peak detector arrangement is useful for providing a simple detection of the bubble point response or for mapping the maximum amplitude of the pulses received at the input. By increasing the resistances of R6 and R7 and decreasing the resistance of R8, the peak detector can be modified to include finite rise and fall times. In this configuration, the conditioning circuit may average the pulses received at the input signal and/or generate an output which tracks the baseline of the input signal.

The threshold comparator 21 may operate to compare the signal output from the conditioning circuit 18 with the threshold voltage. The threshold voltage may be varied to adjust for the parameters of the test setup, such as microphone position, and/or the characteristics of the porous element. Additionally, the level of the threshold voltage may be adjusted depending on, for example, the level of the applied pressure. Whenever the conditioned signal exceeds the threshold voltage, the threshold comparator 21 outputs a threshold detect signal. The threshold detect signal may be input into a variable pulse width detector 46 which detects whenever the threshold detect signal has been continuously activated for a predetermined time period. The predetermined time period is adjustable, and, in the preferred embodiment, can be adjusted between about 0.01 and about 1.0 second. The variable pulse width detector 46 outputs a pulse width detect signal whenever the pulse width- of the threshold detect signal exceeds the predetermined time period. The pulse width detect signal preferably activates the threshold indicator 22 and provides a visual indication whenever the threshold voltage is exceeded for the predetermined time period.

The conditioning circuit 18 and the threshold comparator 21 provide a means for measuring when the average sound pressure level continuously exceeds a threshold level for a predetermined period of time. Detecting the sound pressure level may be implemented in other ways, including, for example, such circuits as an RMS circuit or a low pass filter/integrator circuit.

When an ultrasonic test is initiated, the stabilizing indicator 25 may turn on and remain illuminated for a stabilizing time period determined by the stabilizing time set switch 31. Typically, the stabilizing time set switch 31 is set to indicate a time of about 15-20 seconds. During this time the housing 2 is pressurized and the wetted porous element stabilizes. During the stabilization period, the signal from the threshold comparator 21 is ignored. Referring to Figure 7, following the start input (Figure 7A) , the stabilizing indicator 25 is illuminated (Figure 7B) . The stabilize indicator 25 remains on for the time necessary to increase the pressure in the housing 2 and allow the microphone response to stabilize (stabilization time) . An exemplary simplified microphone response is shown in Figure 7F, with the stabilizing period specifically identified. The stabilizing indicator 25 turns off after the filter has stabilized (typically 15-20 seconds) , and the testing indicator 26 may then be illuminated. The testing indicator 26 remains on until the test is complete as determined by the total test time set switch 28. Typically, the testing indicator 26 remains on for a period of about 45-50 seconds. During this period, the threshold detector signal may be examined by the output logic 27.

A porous element which is not defective may nevertheless have a response that includes a number of noise spikes. These noise spikes were found to be present in non-defective porous elements and are believed to be due to a variety of causes, including: (a) liquid dripping from the porous element 3 into the housing 2, into the outlet tube 6, and/or onto the microphone 4; (b) liquid moving on the surface of the porous element, the housing, and/or the microphone, (c) external acoustical noise from outside of the test set-up, (d) self generated electrical noise caused by the electronics of the signal processing device 9, and/or (e) bubbles on the surface of the porous element.

Several measures have been found to reduce this noise and increase overall system sensitivity including: a) increasing the signal to noise ratio for the microphone 4, pre-amplifier 15, bandpass filter 16, and variable gain amplifier 17; b) incorporating a high Q bandpass filter optimized to the optimum mechanical to electrical efficiency frequency of the transducer; c) shielding the microphone 4 from external acoustic noise, e.g., by placing the microphone 4 inside of the elastomeric fitting 43, the housing 2, the inlet and/or the outlet tubes; d) shielding the preamplifier 15 from external electrical noise, e.g., by placing the preamplifier 15 inside of a conductive enclosure capable of shielding the preamplifier 15 from external electromagnetic radiation and by reducing the electromagnetic radiated emissions from other portions of the circuit; e) pre-pressurizing the assembly at half the anticipated bubble point test pressure before increasing to the full test pressure (note that small porous elements do not require this pre- pressurization procedure since the pressurization time necessary to remove excess wetting solution is substantially inversely proportional to the surface area of the porous medium) ; f) vacuuming excess wetting solution from the surface of the porous medium; g) shielding the microphone from any contact with the wetting solution by, for example, placing the microphone downstream from the drain; h) electro-polishing the surface of the microphone to form a hydrophobic surface; i) incorporating a circuit to de-pressurize the housing 2 before it is opened after the test, preventing the high pressure inlet side from blowing wetting solution into the low pressure outlet side and onto the microphone surface; j) removing any liquid from the surface of the microphone; k) using separate analog and digital power supplies or voltage regulators to prevent coupling through the power supply, eliminating ground loops to prevent coupling between the digital and analog portions through the common supply, and using slow CMOS to reduce current modulation; and 1) coupling the microphone 4 to the preamplifier 15 using shielded cable. Despite these efforts, it has been found that some noise spikes or glitches still remain. Consequently, the signal processing device incorporates an arrangement such as the conditioning circuit 18, the pulse width detector 46, a pulse counter and/or other suitable circuitry to discriminate noise spikes present in a non-defective porous element from the noise made by a defective porous element.

For example, a microphone response for a hypothetical defective porous element is illustrated in Figure 7F. Whenever the pulse width detect signal indicates that the threshold voltage has been continuously exceeded by the conditioned signal for more than a predetermined noise limit period (typically 0.01 - 1.0 seconds), a failure may be indicated by illuminating the fail indicator 23, as shown in Figure 7G. If the pulse width detect signal indicates that the threshold voltage has not been continuously exceeded by the conditioned signal for the noise limit time period (typically 0.01 - 1.0 seconds) , the pass indicator 24 may be illuminated.

Throughout the ultrasonic test, the master counter 34 may increment to mark the test time. The master counter 34 typically increments from 1 to about 45-50. The pass indicator 24 or the fail indicator 23 remains on until the start of the next test cycle. In order to calibrate the system response with the physical characteristics of a particular filter to be tested, it may be preferable to adjust the noise limit time, the stabilizing time, the total test time, the bandwidth of the bandpass filter 16, the gain of the variable gain amplifier 17, the rise and fall time of the conditioning circuit 18, the threshold voltage of the threshold detector 21 and the pulse width detected by the pulse width detector 46. However, in some production environments, it may be preferable to fix these values for a single type of porous element and/or test system configuration.

Figure 9 illustrates another method for testing a porous element. In this method, a first test is run as described above. This allows the porous element to be characterized as meeting a minimum standard for integrity. Following the first test, a second test is performed to determine the bubble point of the porous element. In the second test the pressure in the housing 2 is slowly increased until the pulse width detect signal indicates that the threshold voltage has been continuously exceeded by the conditioned signal for more than a predetermined noise limit period, indicating that the bubble point has been reached. The pressure at which the porous element reached its bubble point may then be used to calculate the maximum pore diameter for the porous element using well known methods. The second embodiment allows different porous elements to be classified into different grades. It also allows for a quantitative measure of pore size. Pore size characterization is desirable because porous elements may have a bubble point which is too high due to the wrong material, defective materials, or a clogged medium. The measured bubble point can be reported, collected, and analyzed as a check on the manufacturing process. A sensitivity adjustment is provided which allows the signal processing device 9 to tune from full sensitivity, for example, for ultra porous elements to a reduced sensitivity for macro porous elements. The sensitivity adjustment is typically accomplished by adjusting a combination of the gain on the variable gain amplifier 17 and the threshold voltage value provided to the threshold comparator 21.

Low surface tension liquids can be used with the porous element testing system 1. Because the porous element testing system 1 does not rely on measuring the leak to diffusion ratio, it is relatively unaffected by low surface tension fluids. Low surface tension fluids have the advantage of being superior agents for wetting the porous elements (especially hydrophobic porous elements) and allowing the porous elements to dry quicker once the test has been completed.

In many applications, it may be desirable to perform both forward flow tests, which are well known in the art, and ultrasonic tests embodying the invention on the same porous element. It has also been found to be desirable to perform ultrasonic testing in both the forward and reverse direction. Performing both a forward flow test and ultrasonic tests on the same porous element has been found to provide improved reliability. It has been determined that it is even more effective to conduct both tests simultaneously.

The porous element testing system 1 can be installed in existing forward flow test stands or in on-line testing applications with minimal modification. For on-line applications, it is now practical for a end-user to conduct on-line reverse bubble-point testing of filters without the requirement to visually observe the filter. Conventional bubble point tests required an operator to observe the filter or employ an ultrasonic detector coupled to the filter via a liquid medium. The liquid medium is extremely disadvantageous because the ultrasonic transducer is coupled to noise sources originating down-stream from the filter-under-test as well as noise sources originating in structures adjacent to the fluid filed piping. The signal to noise level masks the defective filter signals. Additionally, during visual reverse bubble point testing, an operator can distinguish bubbles resulting from diffusional flow from bubbles caused by defects by prodding the filter with a probe and watching whether the source location of the bubbles moves. However, in on-line testing or automated testing using fluid coupled ultrasonic sensors, it is frequently not possible to distinguish between diffusional flow and true leaks. Because of these problems, conventional fluid coupled ultrasonic testing arrangements are typically limited to laboratory environments, and require extensive insulation. Even with these measures, conventional fluid coupled ultrasonic testing apparatuses are limited to identifying gross sound levels such as the large increase in sound levels that occur at the bubble point. By contrast, embodiments of the present invention are not as susceptible to environmental noise and provide greatly increased discrimination between defective and non- defective filters. It has been found that the use of air or other gaseous phase fluids to couple the microphone to the porous element better insulates the microphone from extraneous noise sources, while still permitting the microphone to detect sounds produced by the filter medium. Thus, the gaseous coupled ultrasonic testing makes reverse bubble point testing in operational environments possible.

Forward flow testing and ultrasonic testing can be performed simultaneously, and, generally, performing both tests simultaneously does not require any more time than the time required by a single forward flow test. The housing 2 is preferably pressurized to the same pressure as specified for a forward flow test. The microphone 4 can be placed in the downstream part of the outlet tube 6 and left there, even during sterilization. This is possible because the preferred microphone 4 is capable of operating after exposure to temperatures in excess of 300^βF.

Another embodiment of the ultrasonic leak detector employs a sonic test method for assessing the integrity of porous elements. The test method is termed the "pulse count test" and measures the pulse density of sound pressure levels generated by a wetted porous element as a gas pressurizes the up-stream surface of the porous element. It has been found that this testing method is a very effective tool in discriminating properly functioning filters from defective ones.

Wetted filters under pressure produce a number of sound pressure pulses, superimposed on a baseline signal level. Figure 9 shows a typical response of a filter where the sound pressure has been converted to electrical energy by the microphone 4, amplified, filtered, and conditioned. The graphical representation shown in Figure 9 is a simplified response. Detailed graphs of the response of typical non-defective and defective filters are respectively shown in Figures 13 and 14.

Referring to Figure 13, the microphone response of a non-defective filter is shown as the applied pressure is increased to a level below the bubble point. As the initial pressure is increased, the forward flow curve shows a higher initial flow followed by a stable flow rate. The higher initial flow results from, for example, flexing of the filter in the down stream direction in response to the applied pressure. The period of time required to reach a stable flow is termed the stabilization time. Similarly, the sound level within the test chamber increases initially as the porous element is pressurized, followed by a relatively stable sound pressure level. Much of the signal activity during the relatively stable portion of the sound pressure level period in Figure 13 is attributable to noise. The pulse counts per unit time are shown at the bottom of Figure 13.

Referring to Figure 14, the microphone response of a defective filter is shown as the applied pressure is increased to a level below the specified bubble point of the filter. Various indicators can indicate that a filter is defective. The sound pressure level of the filter during the relatively stable portion of the sound pressure curve is substantially greater in amplitude, on the average, and in pulse density as compared to the sound pressure curve of Figure 13. The defective filter can be determined from a measure of the amplitude of the pulse (peak detection) , a measure of the average energy of the pulses (average energy) , a measure of the pulse density, and/or by other suitable mechanisms for measuring sound signal levels and variations.

As the applied pressure reaches the bubble point, or at pressures below the bubble point for defective filters, the number of pulses per unit time (pulse density) increases and the amplitude of the pulses increases. As shown in Figures 9, 13, and 14, as the pulses density increases, the pulses begin to merge together. The next pulse begins before the microphone and electronic signal processing circuits have stabilized from the previous pulse. Under these circumstances, the baseline of the pulses rises. By conditioning (averaging or peak detecting) the pulses or groups of closely spaced pulses, the rise in the baseline, for example, can be detected. This information can also be utilized to discriminate between defective and non-defective filters and to determine the bubble point of a filter as described above.

Alternatively, it has been found that by counting the pulses directly, a greater level of discrimination between defective and non-defective filters can be achieved for many porous elements.

The pulse density can be measured directly by counting the number of pulses that occurred within a particular interval in time. The time intervals may be selected to be every fraction of a second, every second, every two seconds, etc. For example, the number of pulses that occurred in the 1st and 2nd seconds after stabilization could form a first pulse count, the number of pulses that occurred in the 3rd and 4th seconds after stabilization could form a second pulse count, etc.

It was found that this technique has a problem in that if a particular cluster of pulses occurs during, for example, the 2nd and 3rd second, the pulse count will be divided between two time intervals and a loss of sensitivity will result. One method of overcoming this loss of sensitivity is to utilize a sliding window technique to measure the pulses. For example, a first pulse count counts all pulses in the 1st to 5th seconds after stabilization, a second pulse count counts all pulses in the 2nd to 6th seconds after stabilization, etc. The sliding window is not limited to operation in increments of one second. Depending on the particular apparatus employed to perform the pulse counting, it is possible to determine the pulse density for any size window with any size increment. The sliding window technique may be performed so that each window has a predetermined duration, and each subsequent window overlaps a previous window by a predetermined amount. If extreme precision is required, a new window pulse count can be calculated each time that an additional pulse is received. It has been found that the use of a sliding window for measuring pulse density provides better discrimination of defective filters than the direct pulse counting per unit time. Regardless of whether the pulses are counted directly or using a sliding window, a failure is indicated if the measured count exceeds a predetermined limit, ϊhis limit may be a fixed limit or vary with increasing pressure.

The pulse counting technique may be implemented using a plurality of different apparatuses. Figure 10 shows a first exemplary embodiment of a pulse counting apparatus. In the apparatus shown in Figure 10, a sliding window pulse counter is implemented where each window has a duration of five times the clock frequency and each subsequent window overlaps the previous window by four times the clock frequency. For example, where a clock frequency of 1 Hz is utilized, then the apparatus shown in Figure 10 has an individual window duration of 5 seconds and each subsequent 5 second window overlaps the previous window by 4 seconds. An amplified signal from a microphone may be input into a smoothing circuit 58. The amplified signal, may, for example originate from the output of the variable gain amplifier 17 as shown in Figure 3. The smoothing circuit 58 is optionally provided for smoothing the pulses to better match the response time of other components in the pulse counting circuitry 60. One example of a suitable smoothing circuit 58 is shown in Figure 4A. The amplified signal (with or without smoothing) may be input into a first input of a comparator 61.

The second input into the comparator may be a predetermined threshold voltage V_τhreshold2 or a varying threshold voltage V_{Baββ ne} that varies with the baseline of the microphone signal received depending on the setting of a mode select switch 59. The predetermined threshold voltage V_τhreshold2 provides a constant DC voltage at a predetermined level. The predetermined level may be varied to adjust the for such variables as microphone position, housing type, or the characteristics of the porous element. The varying threshold voltage V_Baseline may, for example, originate from the conditioning circuit 18 of Figure 3 in those embodiments where the conditioning circuit is implemented by a circuit having relatively long rise and fall times. For example, the varying threshold voltage V_Baseline may be obtained from the conditioning circuit 18 when the conditioning circuit is implemented using the circuit shown in Figure 4B and when the resistances R6-R8 are adjusted so that the conditioning circuit 18 tracks the baseline of the amplified signal.

The output of the comparator 61 may be input into a one-shot 62. The one-shot 62 outputs a square wave pulse responsive to pulses received from the comparator 61. The duration of the pulse output from the one-shot 62 is not critical so long as the duration of the pulse is not so long as to mask subsequent pulses. If the porous element produces pulses at a rate of, for example, 100 Hz, then the one-shot 62 is preferably set to have a pulse duration of about 1 milli-second. The output from the one-shot 62 may be input into a single counter for counting the pulses directly, or input into a plurality of counters for implementing the sliding window counting method. In a preferred embodiment of the exemplary apparatus shown in Figure 10, the output from the one-shot 62 is input into a clock input of counters 63-67. The counters utilized in the pulse count circuitry 60 may comprise any suitable counting mechanism and include any number of counters. In the preferred embodiment, there are 5 counters (counters 63-67) , each comprising 8-bit decimal counters for counting between 1 and 100. The 8-bit output from each of the counters 63-67 may be input into a bus multiplexer 72. The bus multiplexer may be implemented, for example, using a plurality of data selector circuits such as a MC14512 manufactured by Motorola. The bus multiplexer 72 may selectively output the results from 1 of the 5 counters to a latch circuit such as 8-bit latch 73. The results contained in the 8-bit latch 73 may be output to a plurality of output devices for analyzing the results of the test such as a display 75, a D/A converter 77, and/or a comparator 76. The display 75 may be any display capable of visually displaying the current value of the latch 73. The comparator 76 preferably receives an adjustable count limit indicative of the maximum number of pulse counts permitted for a particular porous element. It may be desirable to visually indicate a failure whenever the output from the latch 76 exceeds the count limit. Suitable logic for visually indicating that the pulse count has been exceeded may, for example, be constructed as shown in Figure 6 with the output from the comparator input into the pulse width exceeded signal. The D/A converter 77 may be coupled to a chart recorder for producing a plot of the pulse count as shown in Figure 11.

Timing and control for the pulse count circuitry 60 may be provided by any suitable mechanism. In the illustrated embodiment shown in Figure 10, the timing and control is provided by a clock signal CLK input into an inverter 74, down counter 68, decoder 69, timing control NAND gates 70, and master reset NAND gates 71. The clock signal CLK may be operated at any frequency. In the preferred embodiment, the clock is operated at a frequency of 1 Hz.

In operation, the embodiment illustrated in

Figure 10 provides a sliding window pulse counter where each window has a duration of five times the clock frequency and each subsequent window overlaps the previous window by four times the clock frequency.

With a 1 Hz clock, this provides a five-second window with one second increments. The apparatus counts the number of pulses in the 5-second window, and updates this count at intervals of one second. The digital display shows the count in the latest 5-second interval, and is updated every second. A failure is indicated if the measured count exceeds a preset limit. In addition, a D/A Converter provides an analog output, suitable for a strip chart recorder. This provides a permanent record, in a graphical bar chart format, of the complete test.

When the mode select switch is set to the threshold voltage ^v _τhreshoi_d2' pulses received which exceed the threshold voltage V_τhreshotd2 trigger the one- shot 62 to produce a pulse. The pulses from the one- shot 62 increment each of the counters 63-67 simultaneously. Reset inputs Reset 1-5, respectively coupled to counters 63-65, provide asynchronous resets to each of the counters so that a different counter is reset on each cycle of clock input CLK. The timing and control circuitry operates so that the rising edge of the clock input CLK controls the multiplexer 72 to select a new counter to output to the latch. The falling edge of the clock input CLK latches the current count of the selected counter into the latch and resets the selected counter. Counting in the reset counter then resumes from an initial value of zero. The down counter 68 sequentially selects each of the counters so that the counters are output and reset in the order R5, R4, R3, R2, Rl, R5, ... If a 1 Hz clock is input into the input clock CLK, then each counter will be reset every 5 seconds, and a different counter is reset every second. In this manner, the sliding window described above is implemented.

A sample output from the chart recorder 78 is shown in Figure 11 for a response to a bubble point such as the one shown in Figure 9 for the case where the threshold voltage is set to a fixed constant. If, for example, the predetermined count limit input into the comparator 76 had been set at 16 pulses (in 5 seconds) , a failure would be indicated at the 10-second point.

Referring to Figure 9, when the voltage threshold ^vτ_hreshoi_d2 ■"-^{s set to a f}i^χed constant and the one-shot 62 is edge triggered, there will be a large number of pulses prior to reaching the bubble point and then the pulses will cease. This is illustrated, for example, in Figure 11. Pulse counting ceases because the baseline of the signal response from the microphone exceeds the threshold voltage V_τhreshold2 so that the one- shot 62 is no-longer activated (assuming a one-shot 62 that is edge triggered) .

Other circuit arrangements are possible which continue to count pulses even after the bubble point is reached. For example, when the mode switch in Figure 10 is set to the threshold voltage V_Baseljne, the threshold voltage tracks the baseline. In this arrangement, pulse counting continues even after the baseline voltage increases because the threshold voltage V_Baβeline is configured to track the baseline. Other circuits can be constructed to provide similar results. For example, a circuit capable of storing the average voltage may be suitable for providing the threshold voltage V_BasβUne.

Although the strip chart recorder 78 provides a useful diagnostic tool, for production testing, only a simple limit setting is required so that there is no interpretation of data by the operator. This is a significant advantage over conventional bubble point testing where operator experience and judgement was a factor in reverse bubble point testing.

A fourth embodiment of the porous element testing system 1 is illustrated in Figure 12. Components in the fourth embodiment are similar to components in the other embodiments. In the fourth embodiment, the housing 2, containing the porous element, is coupled to a forward flow meter 55, a gas control system 54, a microphone 4, a transducer 53 (coupled either to the inside or the outside of the housing 2) , and sensors 57. The microphone 4 detects sound within the housing 2 and outputs a signal indicative of the sound detected.

The signal from the microphone is amplified by the preamplifier 15 which outputs a preamplified signal. The preamplifier 15 is preferably constructed as discussed above with respect to the first embodiment. Alternatively, the preamplifier 15 may be constructed using a standard ultrasonic preamplifier. In some filtration devices, it may be preferable to include a plurality of microphones 4. The plurality of microphones may be arranged adjacent to a single porous element and coupled together to increase the signal to noise (S/N) ratio. The random noise associated with, for example, electrical noise of the electronic components will not be additive, while sound signals from the filters will be additive, thus increasing the signal to noise ratio. Any method for in-phase coupling the signals from the two microphones may be utilized as is well known in the art. For example, the signals may be added using an analog adding circuit disposed in the system either prior to or after the preamplifier 15. Alternatively, the signals from the microphone may be added digitally. In this arrangement, a plurality of microphones 4 are respectively coupled to sperate channels, each channel containing a preamplifier 15 and A/D converter 45 coupled to the signal processor 49. The signal processor may then add the response of each channel digitally.

Conditioning circuit 47 may optionally be provided to provide amplification, filtering, and/or signal conditioning as discussed above relative to the other embodiments. In a preferred embodiment, filtering is provided to limit the bandwidth of the preamplified signal to, for example, the ultrasonic range prior to digital conversion by an A/D converter 45. Alternatively, the A/D converter 45 can receive the signal directly from the pre-amplifier 15. The A/D converter converts the received signal into a digital signal. The digital signal is input into a digital signal processor 49. The digital signal processor 49 may, for example, be a dedicated signal processing device or a programmable computer such that the operator display terminal 50 and the signal processor 49 are combined. In a preferred embodiment, the operator display terminal 50 and the signal processor 49 are implemented in a single programmable computer using LabVIEW For Windows software from National Instruments. Figure 16 is a graphical representation of one embodiment of a LabVIEW For Windows software program for controlling the signal processor 49.

The digital signal processor 49 is coupled: to the forward flow meter 55 for measuring the forward flow of a gas through the housing 2, to the gas control system 54 for controlling the pressure within the housing 2, to an operator display terminal 50 for providing control and input data, to a digital to analog (D/A) converter 51 for outputting built-in-test (BIT) signals, and to a plurality of sensors for monitoring the environment in which the test is conducted. The digital to analog converter 51 is coupled to a voltage controlled oscillator/amplifier 52, which is in turn coupled to transducers 53.

In a preferred embodiment, the transducer 53 is coupled to the outside of the housing 2. Sound imparted to the outside of the housing 2 may be detected by the microphone 4 located on the inside of the housing 2. By coupling the transducer 53 to the outside of the housing 2, the transducer 53 is insulated from the fluid processed by the porous element testing system 1. In a preferred embodiment, the transducer 53 is preferably be an ultrasonic transducer and may be the same type of ultrasonic transducer utilized for the microphone 4. Alternatively, the transducer 53 may be a audio speaker, or any other mechanism for converting energy into sound. In an alternative embodiment, the transducer 53 may be coupled to the inside of the housing 2 to provide additional sensitivity for the built-in-test.

It is also possible to couple the VCO/amplifier 52 directly to the microphone 4 so that the microphone 4 serves both as both a transducer and a microphone. In this manner, a pulse can be originated by the VCO/amplifier 52, converted to an ultrasonic sound wave by the microphone 4 producing an incident sound wave within the housing 2. A reflected sound wave may then be received by the microphone 4, converted into an electrical signal and received by pre-amplifier 15. Thus, a BIT may be conducted by a single ultrasonic transducer. However, this method is less preferred since by using a single transducer for both transmit and receive functions, the sensitivity of the transducer for receiving sound signals is reduced.

The signal processor 49 controls the A/D to initiate a built-in-test signal. The voltage controlled oscillator (VCO)/amplifier 52, may for example, be constructed using two cascaded precision wave-form function generators whose output signal is amplified. The precision wave-form generators may be constructed using standard 8038 circuit available from EXAR or Intersil. In a preferred embodiment, a first wave-form generator is set to oscillate at a relatively low frequency such as 100 Hz. The output from the first wave-form generator circuit is used to frequency modulate a second wave-form generator circuit about a fundamental frequency which is set to coincide with the resonance of the transducer 53. In this arrangement, the second wave-form generator circuit outputs a wave-form having a frequency which oscillates between a resonance and non-resonance frequencies in accordance with the output from the first wave-form generator circuit. The output from the second wave-form generator circuit is preferably amplified by an amplifier capable of driving the transducer 53. The transducer outputs a signal in accordance with the frequency of the first wave-form generator circuit.

The signal processor 49 may optionally include a plurality of sensors 57 such as temperature sensors and barometric sensors. Pressure sensors may be useful in providing greater accuracy in the flow meter. Temperature sensors are useful for on-line customer applications where the housing 2 is operated at an elevated temperature due to steam sterilization or process parameters. It is well known that the pores in a porous element may act as capillaries. The pressure required to force fluid through a capillary is related to the viscosity of the fluid flowing through the capillary. Many liquids have a viscosity that varies greatly with temperature. Thus, in order to ensure reliable operation in on-line testing environments having wide temperature variations, it is desirable to monitor the temperature at which the porous element is tested. The test pressure and defective filter parameters may then be adjusted to correspond to a particular viscosity of the wetting fluid. In operation, the embodiment shown in Figure 10 may operate to perform any of the methods and circuit functions hereinbefore discussed for other embodiments of the porous element testing system 1, either individually or in combination. Referring to Figures 13 and 14, the signal processor 49 operates to provide a quantitative measure of the signal produced by the sound transducer 4 in order to discriminate between defective and non-defective porous elements. Quantitative measures produced by the signal processor 49 may include: a) detecting the minimum or maximum peaks of signals (peak detection) to provide a quantitative measure of the peaks of each of the signal pulses; b) detecting the average signal voltage, current and/or power by any suitable averaging technique including rms, maximum or minimum pulse amplitude averaging, integration, low pass filtering, peak detection having finite rise and fall times, and/or signal averaging; c) detecting signal density by any suitable technique including: counting pulses relative to a fixed value or a plurality of differing fixed values, variable average value (baseline) , and/or previous pulse value (differential amplitude pulse counting) ; counting frequency shifts; and/or measuring the time between pulses; and/or d) detecting signal variability by measuring differences in voltage, current, power, and/or frequency.

Each of the quantitative measures can be individually correlated with defective and non-defective filters, and/or processed to provide a confidence index combining a plurality of the quantitative measures. Additionally, each of these quantitative measures can be combined with forward flow measurements and other measurements of filter integrity. In this manner, filters that do not fail individual quantitative measures of integrity but have, for example, values falling at the upper range of a plurality of quantitative measures can be identified for close scrutiny. Additionally, each of the individual quantitative measures can be analyzed statistically to determine the standard deviation, variance, mean, and other statistical attributes. For example, it has been found that the standard deviation for non- defective porous elements may be higher than the standard deviation for non-defective porous element. It has been found that measuring pulses with respect to a plurality of different threshold values provides increased discrimination between defective and non-defective porous elements.

Each of the above mentioned quantitative measures can be calculated for the entire test period, for a particular portion of the test period (quantitative measure per unit time) , and/or for sliding windows where each window has a fixed or variable duration and where windows may overlap previous windows by a fixed or variable amount. The windows may be determined using units of time and/or other measures derived from the signal such a pulse counts or frequency shifts.

For on-line applications, the individual performance of each filter element can be saved from each test. This data provides a history of the response of the filter element under previous tests, and alerts the operator to any substantial deviations from previous tests. If a substantial increase in, for example, the forward flow value or the pulse count value is detected, then the porous element testing apparatus 1 signals the operator that additional off¬ line testing may be desirable.

In one mode of operation, the embodiment of the porous element testing system 1 shown in Figure 12 can be used as a diagnostic tool. This allows certain kinds or sizes of defects to be "finger printed" by their signal characteristics such as frequency. The digital signal processor 49 compares the signal received from the microphone 4 with a number of finger print signatures stored in memory to identify the existence and/or type of filter defect.

Referring to Figure. 8, several examples are illustrated showing various finger prints. The finger prints illustrated in Figure 8 represent the simplified time domain response input into the signal processing device 49 for various conditions of the porous element. Figure 8A shows an example of a test pressurization curve having a 20 second ramp and a 10 second hold period. Figures 8B-8G illustrate simplified drawings of various fingerprints which result from the pressurization curve shown in Figure 8A being applied to porous elements having various conditions. For example, Figure 8B shows the finger print that results when the wetting solution is correctly applied, but the bubble point is too low. Figure 8C shows the case where the wetting solution is correctly applied, and the bubble point is within a permissible range. Figure 8D shows the finger print which results from a bubble point being too high. Figure 8E illustrates the case where an insufficient amount of wetting solution is applied to the porous element, even though the bubble point is within the permissible range. One problem that may be encountered when testing filters using forward flow test methods is that a false indication of a faulty filter may result from insufficient wetting of the filter. By combining forward flow tests with the porous element testing apparatus 1, an improper wetting can be detected, and an operator can be informed that the filter may have been improperly wetted. Under these circumstances, the operator can re-wet the filter and begin the test again. Too little wetting can often be detected by a slowly rising baseline. Figure 8F shows the finger print associated with a bubble point within the permissible range, but where too much wetting solution has been applied. It has been found that is difficult to discriminate a condition of too much wetting from real failures. Therefore, once the signal processor 40 detects that too much wetting may have occurred, it is desirable to have the signal processor 40 either extend the testing time by maintaining the gas control system at the predetermined pressure, e.g., 80% of the bubble point or conduct a second test. Figure 8F shows the finger print associated with a filter having a pin hole. A small pinhole may, for example, appear as small pulses superimposed on the correct output, assuming the porous element is good other than having a pinhole. Some ultrafiltration porous elements have pores that are too small to be tested using bubble point techniques. This may occur when, for example, the bubble point pressure is prohibitively high because of limitations of the test apparatus. These ultrafiltration porous elements typically include a membrane and a structure, such as an end cap, mounted to the membrane. Ultrasonic testing can be used to test for defects in the membrane mounting structure, defects in mounting the membrane to the membrane mounting structure, and gross defects in the membrane. The same finger print techniques described above can be used to classify the various defects in ultrafiltration porous elements as a diagnostic mechanism. Preferably, the digital signal processor 49 classifies the type of defect, and then displays an indication to the user on the operator display terminal 50 indicative of the type of defect detected. In this embodiment, it may be desirable for the band- pass filtering, half-wave rectification, signal integration, threshold detection, pulse width detection, stabilization time, and total test time functions to be performed by the digital signal processor 49. Each of the functions of the digital signal processor 49 are user programmable via the operator display terminal 50. This allows the functions to be tailored to a particular porous element. For example, large porous elements typically have noise spikes of a greater duration. Thus, it is particularly advantageous to be able to vary the predetermined noise limit time period.

Typical outputs from the operator display terminal 50 are shown in Figures 13-15. Figures 13 and 14 are discussed above. Figure 15A shows a graphical illustration of an average of the analog signal sampled at a rate of 20 KHz and averaged using a one second window. The left hand portion of the output screen shown in Figure 15A displays the current analog output voltage, the bubble point K_L determined from the averaged sound signal, and the standard deviation. Figure 15B displays a graphical representation of the pulse count, the 1st bubble point (as determined by the pulse count) , the current pulse count, the standard deviation, and the pulse count error threshold for the particular filter element under test. Figure 15C shows the pressure applied to the porous element under test. Figure 15D displays a graphical representation of the mass flow, the current mass flow, the bubble point K_L as determined from mass flow, and the standard deviation of the mass flow.

The digital signal processor 49 may also be programmed so that it can dynamically set the adjustable bandpass filter's characteristics, variable gain characteristics, pressure level, microphone position selected, total test time, stabilization time, low pass filter parameters, noise limit time and threshold value in response to a code input by the user and indicative of the type of porous element being tested, the type of wetting solution applied to the filter, etc. It should also be noted that the functions of the circuitry illustrated in Figure 3 may be performed directly by the signal processor 49.

When used in conjunction with a forward flow test meter 55, the signal processor is capable of inputting information received from the forward flow test and displaying this information on the display terminal 50. Ultrasonic testing is particularly advantageous for detecting some types of defects such as defects in extremely thin membrane filters that rely on pore size as the filtration mechanism and not on membrane thickness. Additionally, ultrasonic testing is compatible with simultaneous forward flow tests, and with reverse flow testing the porous element. When testing the porous element in the reverse direction, both a gas control system 54 and a forward flow meter 55 can be provided on both sides of housing 2.

It has been found that defects in porous elements occur over a particular pressure range. Thus, in a preferred embodiment, it is desirable to slowly ramp the pressure while monitoring the sonic output from the porous element to increase discrimination between defective and non-defective porous elements. Alternatively, the applied pressure may be stepped in discrete increments to provide increased discrimination and detect defects which occur only at particular pressures.

The digital signal processor 49 also preferably stores a series of test calibration patterns that are associated with certain types of porous element defects to test and calibrate the porous element testing system. These . calibration patterns are utilized to test the microphone 4 and other circuitry to ensure operability and to provide a fail-safe fault detection mechanism that can be actuated, for example, before and after each test sequence.

In operation, the signal processor 49 outputs a digital test signal to the digital to analog converter 51. The digital to analog converter 51 converts the digital calibration signal to an analog calibration signal. The analog calibration signal may, for example, be converted into a signal for driving the transducer 53 by VCO/amplifier 52. Transducer 53 converts the test signal into sound pressure levels within the housing 2. ^"These sound pressure levels are received via microphone 4 and input into digital signal processor 49 via pre-amplifier 15, conditioning circuit 47 and analog to digital converter 48. The signal processor 49 can compare the received test signal with the calibration signal sent and thereby verify the operation of the porous element testing system 1.

The operator display terminal 50 may optionally contain a data base program which automatically receives test data from a particular manufacturing lot of porous elements. In this manner, it can be determined whether an entire lot is within manufacturing specifications when a predetermined percentage of the porous elements under test fail to pass. The operator display terminal 50 may also be utilized by an operator to grade a particular lot of porous elements so as to certify its applicability to a particular type of application based quantitative measurements and the average failure rate of the manufacturing lot.

In certain applications, a plurality of filter devices are tested in parallel. This may occur, for example, in a distillation application where a requirement for a low pressure drop across the filter elements dictates that an extremely large number of filter elements be coupled in parallel. Under these circumstances, it may be difficult to adequately test these filters using forward flow, and further it may be difficult or impossible to isolate a particular faulty filter element from the plurality of filter elements. Using the present test method, it may be possible to include a different microphone in close proximity to each filter element, or group of filter elements. The sound signal form a particular microphone can then be utilized to isolate the failure to a single filter or group of filters.

While the invention has been described in some detail by way of illustration and example, it should be understood that the invention is susceptible to various modifications and alternative forms, and is not restricted to the specific embodiments set forth in the Examples. It should be understood that these specific embodiments are not intended to limit the invention but, on the contrary, the intention is to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the invention. For example, it may be possible to utilize the defective porous element identification methods and apparatuses to detect defects in a fluid coupled sonic testing apparatuses. Additionally, it may be possible to obtain improved data relating to the pore size distribution of the porous element. It is well known that the pore size distribution of porous elements is related to the slope of the vertical portion of the forward flow test curve after the bubble point is reached. Similarly, it may be possible to relate the acoustic signals produced after reaching the bubble point with the distribution of pore sizes.

Claims

What is Claimed is:

1. A porous element testing system for testing a porous element wetted with a wetting solution, the porous element testing system comprising: a housing having a first side and a second side, wherein the first side is divided from the second side by the wetted porous element and wherein the first side and the second side are both filled with a gas; a differential pressure generator for generating a differential pressure across the wetted porous element; a transducer disposed in the vicinity of the porous element for receiving acoustic signals generated within the housing; a signal processing device, coupled to the transducer, for analyzing the acoustic signals received from the transducer and for determining whether the porous element is defective.

2. The porous element testing system as recited in claim 1 including a gas flow meter arranged to monitor gas flow through the porous element.

3. A method for determining whether a porous element is defective comprising: wetting a porous element with a wetting solution; generating a differential pressure between a gas on a first side of the wetted porous element and a gas on a second side of the wetted porous element; monitoring sound levels emanating from the vicinity of the porous element; and determining, as a result of the sound levels, whether the porous element is defective.

4. The method as recited in claim 3, including measuring gas flow through the wetted porous element.

5. A system for testing a wetted porous element, the testing system comprising: a housing having first and second sides, wherein the first side is dividable from the second side by the wetted porous element; differential pressure generator for applying a differential pressure between the first and second sides of the housing; a sound transducer for receiving acoustic signals generated within the housing; and a gas flow meter arranged to monitor gas flow between the first and second sides of the housing.

6. A method for determining whether a porous element is defective, comprising: wetting a porous element with a wetting solution; creating a differential pressure across the wetted porous element; monitoring acoustic signals generated in the vicinity of the wetted porous element; and measuring gas flow through the wetted porous element.

7. A porous element testing system for testing a wetted porous element, the porous element testing system comprising: a differential pressure generator arranged to generate a differential pressure less than the bubble point pressure across a wetted porous element; a transducer disposed in the vicinity of the wetted porous element to receive acoustic signals; a signal processing device, coupled to the transducer, for analyzing the acoustic signals received from the transducer and for determining whether the wetted porous element is defective.

8. The porous element testing system as recited in claim 7 wherein the differential pressure generator generates a differential pressure by applying a gas having a first pressure to a first surface of the wetted porous element and a gas having a second pressure different than the first to a second surface of the wetted porous element.

9. The porous element testing system as claimed in claim 8 including a gas flow meter arranged to monitor gas flow through the wetted porous element.

10. The porous element testing system as claimed in claim 7 including a means for inducing a sonic signal detectable by the transducer.

11. A method for determining whether a porous element is defective comprising: wetting a porous element with a wetting solution; exposing first and second sides of the porous element to a differential pressure less than a bubble point pressure; monitoring sound levels adjacent to the porous element; determining, as a result of the sound levels, whether the porous element is defective. 12. The method as recited in claim 11, including measuring gas flow through the wetted porous element.

13. Apparatus for testing a wetted porous element comprising: a transducer positionable in the vicinity of the wetted porous element to receive acoustic signals and a signal processing device coupled to the transducer to count pulses in the acoustic signal and determine whether the porous element is defective in accordance with the pulse count.

14. A method for testing a wetted porous element comprising: counting acoustic pulses emanating from the wetted porous element and determining whether the porous element is defective in accordance with the pulse count.

AMENDED CLAIMS

[received by the International Bureau on 4 May 1994 (04.05.94); original claim 2 amended; new claims 3-5 and 16-35 added; claims 3-12, 13 and 14 renumbered as claims 5-15, 36 and 37; other claims unchanged (7 pages)]

1. A porous element testing system for testing a porous element wetted with a wetting solution, the porous element testing system comprising: a housing having a first side and a second side, wherein the first side is divided from the second side cy the wetted porous element and wherein the first side and the second side are both filled with a gas; a differential pressure generator generating a differencial pressure across the wetted porous elemen ; a transducer disposed in the vicinity of the porous element and receiving acoustic signals generated within the housing; a signal processing device, coupled to the transducer, analyzing the acoustic signals reoeived from the transducer for determining whether the porous element is defective.

2. The porous element testing system of claim l including a gas flow meter monitoring gas f ow through the porous element.

. The porous element testing system of claim 1 including a mechanism shielding the transducer from contact with the wetting solution.

. The porous element testing system of claim 1 including a mechanism for reducing the differential pressure prior to opening the housing.

5. The porous element testing system of claim l wherein the differential pressure generator generates a first differential pressure having a first pressure in the f rst side larger, than a second pressure in the second side and a second differential pressure having a third pressure in the first side less than a fourth pressure in the second side, the signal processing device analyzing the acoustic signals received from the transducer while the porous element is exposed to the first and second differential pressures.

6. A method for determining hether a porous element is defective comprising: wetting a porous element with a wetting solution; generating a differential pressure between a gas on a first side of the wetted porous element and a gas on a second side of the wetted porous element; monitoring sound levels emanating from the vicinity of the porous element; and determining, as a result of the sound levels, whether the porous element is defective.

7. The method of claim 6 including measuring gas flow through the wetted porous element.

β. A system for testing a wetted porous element, the testing system comprising: a housing having first and second sides, wherein the first side is dividable from the second side by the wetted porous element; differential pressure generator creating a differential pressure between the first and second si es of the housing; a sound transducer receiving acoustic signals generated within the housing; and a gas flow meter arranged to monitor gas flow between the first and second sides σf the housing. 9. _A method ar determining whetner a porous element is defective, comprising: wetting a porous element with a wetting solution; creating a differential pressure across the wetted porous element; monitoring acoustic signals generated in the vicinity of the wetted porous element; and measuring gas flow through the wetted porous element.

10. A porous element testing system for testing a wetted porous element, the porous element testing system comprising: a differential pressure generator arranged to generate a differential pressure less than the bubble point pressure across a wetted porous element; a transducer disposed in the vicinity of the wetted porous element to receive acoustic signals; a signal processing device, coupled to the transducer, for analyzing the acoustic signals received from the transducer and for determining whether the wetted porous element is defective.

11. The porous element testing system of claim 10 wherein the differential pressure generator generates a differential pressure by applying a gas having a first pressure to a first surface of the wetted porous element and a gas having a second pressure different than the first to a second surface of the wetted porous element.

12. The porous element testing system as claimed in claim 11 including a gas flow meter arranged to monitor gas flow through the wetted porous element. 13. The porous element testing system as claimed in claim 10 including a means ±sr inducing a sonic signal detectable by the transducer.

14. A method for determining whether a porous element is defective comprising: wetting a porous element with a wetting solution; exposing first and second sides of the porous element to a differential pressure lβ3s than a bubble point pressure; monitoring sound levels adjacent to the porous element; determining, as a result of the sound levels, whether the porous element is defective.

15. The method of claim 14 including measuring gas flow through the wetted porous element.

16. The method of claim 14 including inducing a sonic signal detectable by the transducer.

17. The method of claim 14 wherein determining whether the porous element is defective includes setting a discrimination threshold and registering variations in the sound levels that exceed the discrimination threshold.

18. The method of claim 17 including counting variations in the sound level that exceed the discrimination threshold as pulses.

13. The method of claim 18 including counting the number of pulses that occur over a predetermined time period. 20. The method of claim 16 including counting the number of pulses that occur over a predetermined time period, the predetermined time period being determined in accordance with a sliding window technique.

21. The method of claim 17 wherein determining whether the porous element is defective includes setting a plurality of discrimination thresholds having differing levels.

22. The method of claim 17 wherein determining whether the porous element is defective includes varying the discrimination threshold level.

22. The method of claim 22 including counting variations in the sound level that exceed the discrimination threshold as pulse.

24. The method of claim 13 wherein determining whether the porous element is defective includes measuring the time between pulses.

25. The method of claim 14 wherein determining whether the porous element is defective includes utilizing statistical measures to discriminate between defective and non-defective filters.

26. The method of claim 25 wherein utilizing statistical measures includes utilizing a measure of standard deviation of the sound pressure levels.

27. The method of claim 25 wherein utilizing statistical measures includes utilizing a measure of variance of the sound pressure levels. 28. The method of claim 14 wherein determining whether the porous element is defective includes utilizing an averaging technique to average values of a signal produoed by sound pressure levels.

29. The method of claim 14 wherein determining whether the porous element is defective includes utilizing a measure of peak sound pressure levels.

30. The method of claim 14 wherein determining whether the porous element is defective includes utilizing a measure of minimum sound pressure levels.

31. The method of claim 14 wherein determining whether the porous element is defective includes measuring widths of pulses in the sound pressure levels.

32. The method of claim 14 wherein determining whether the porous element is defective includes measuring amplitudes of pulses in the sound pressure levels.

33. The method of claim 14 wherein determining whether the porous ^'element is defective includes measuring a signal density.

3 . The method of claim 14 wherein determining whether the porous element is defective includes measuring frequency shifts.

35. The method of claim 14 wherein determining whether the porous element is defective includes measuring sound pressure level variability. 36. Apparatus for testing a wetted porous element comprising: a transducer positionable in the vicinity of the wetted porous element to receive acoustic signals and a signal processing device coupled to the transducer to count pulses in the acoustic signal and determine whether the porous element is defective in accordance with the pulses counted.

37. A method for testing a wetted porous element comprising: counting acoustic pulses emanating from the wetted porous element and determining whether the porous element is defective in accordance with the acoustic pulses counted.