Background to the Invention

Sealed packaging is used with degradable items and offers protection from moisture, air and biological attack. The most common examples of degradable items are food products, which are packaged in sealed containers, for example, vacuum formed packing or flow wrapped packaging. The seals formed in these processes can be defective due to changes in the process, materials, trapped product debris, machine failure and contamination, the result being that air, moisture and/or bacteria can enter the packaging causing damage to the product.

There are a number of current methods for detecting leaks in packaging and each of these methods suffer from distinct disadvantages. Carbon dioxide leak testing involves a product being packed in carbon dioxide and a sensor seeks out increased concentration of the gas due to a leak. Helium detectors look for helium leaking from the packaging and this gas would normally have to be added to the packaging for the purposes of leak detection. A bubble test requires the requires the packaging to be immersed in water under a vacuum and resulting air loss from the leaks can be seen as bubbles in the water.

It would be desirable to provide improvements in relation to leak detection. Summary of the Invention

A first aspect of the invention provides an apparatus for detecting a leak as claimed in claim 1. The apparatus is particularly suited for use with sealed packages, for example food packages. Typically, such packages comprise a container and/or a sleeve having at least one sealed interface. At least part of the packaging is formed from a material that is flexible. This allows the packaging to deform, and in particular to inflate, in response to changes in pressure. At least part of the packaging may be formed from an elastic material that allows it to stretch in response to changes in pressure.

Advantageously, the apparatus is arranged to removably receive at least a respective one of a plurality of sensor heads, each sensor head comprising a respective different type of sensing device. Said sensing devices may include a plethysmographic sensing device (for example a sensing device arranged to detect changes in electrical impedance or mechanical distension); a force sensor (for example a load cell) and/or a displacement sensing device (for example a capacitive sensor or an electromechanical inductive sensor). Alternatively, said sensing device may comprise any suitable transducer that is capable of detecting changes in said package, or other item, that occur in response to changes in the level of vacuum in the chamber.

Plethysmographic sensing devices are typically configured for measuring changes in volume. In the present case, the sensing devices may be configured to measure changes in the volume of the package under test. Advantageously, this is achieved by measuring one or more characteristics of the surface area of the packaging, in particular changes in the size of the surface area and/or changes in one or more electrical property (e.g. electrical impedance) of the packaging that results from a change in the size of the surface area, especially when the size of the surface area changes elastically under the influence of the vacuum. To this end, an electrical impedance sensor may be used. Spaced apart electrodes (typically two electrodes) may be placed in contact with the packaging and electric current is passed between them via the packaging.

Corresponding changes in the electrical impedance occur as the volume of the package changes, and so the measured electrical impedance can be used as an indicator of package volume. Alternatively, or in addition, a sensor comprising one or more length of elastic and electrically conductive material may be used, whereby the electrical sensor output is determined by the length of the elastic material. The elastic component typically takes the form of a band. The band may be placed against the packaging such that its length changes in response to changes in the size of the surface area of the packaging,

Typically, the band is located around the circumference of the packaging. Hence, as the volume of the packaging changes, the band stretches or contracts correspondingly to produce an output signal that is indicative of the change in package volume (and so the sensor measures mechanical distension of the package). A silastic sensor, or a silastic strain gauge, is an example of a suitable elastic band sensor. Typically, the elastic component of a silastic sensor comprises a tube of silicone rubber filled with a semiconductor material, e.g. indium-gallium. An advantage of the aforementioned plethysmographic type sensors is the ability to detect distension around the circumference of a package, rather than detecting force along a single axis. Preferably, said plethysmographic, or other, sensors are arranged to measure elastic stretching of the package and/or non-elastic inflation of the package (which may occur prior to elastic stretching).

Preferably, said removable sensing head is carried by an extendible device, e.g. an adjustable extension arm, that is operable to accommodate packages, or other target items, of differing heights.

Advantageously, the sensor head includes means for constraining a portion of the package, or other target item, said constraining means defining a perimeter around said sensing device such that, in use, engagement of said contacting means with said package, or other item, defines a portion of said package, or other item, with which said sensing device is operable. The constraining means may for example comprise an annular contact member, e.g. collar, or a plurality of contact members, e.g. blocks, (which may or may not be contiguous) arranged to define a perimeter around all or part of the item under test. Typically, the constraining means constrains the surface area of packaging, or other item, that is affected by the applied vacuum and co-operable with the sensing device in order to amplify the effect being measured by the sensor head. This allows the impact of the external vacuum to be localised to a smaller region inside the constraining perimeter. In a typical implementation a constraining device, preferably an annular constraining device, is pressed against the package, holding the packaging material against its contents, and only permitting the external vacuum to impact on the portion of the packaging that falls inside the constraining device around the sensor. The constraining means is particularly suitable for use with a load cell.

In cases where the sensing device comprises a plethysmography sensor arranged to detect distension changes, the sensing device may comprise a silastic sensor, for example an indium-gallium silastic sensor, or other distension detecting means.

In cases where the sensing device comprises means for measuring mechanical force, it may comprise for example, a load cell, a piezo-electric sensor or an electromechanical lever arm. The extendible member for carrying the sensor head preferably is mounted anywhere within the vacuum chamber. Preferably, the sensor head is carried by a mechanism that is movable (preferably at least vertically, but optionally also horizontally) to locate the sensor head at different places within the chamber.

In typical embodiments, the apparatus includes a plurality of sensor heads, each of which may be removable or not. The sensor heads are arranged in an array, for example a one dimensional linear array, to facilitate testing a plurality of packages, or other items, simultaneously, for example when on a conveyor.

In preferred embodiments, the or each sensor head is coupled to a control unit for receiving and processing output signals generated by the sensor(s) during use. Typically, the control unit comprises a suitably programmed processor, e.g. microprocessor or microcontroller. Advantageously, the control unit includes, supports or is co-operable with one or more of the following: a graphical user interface, operating software, data storage memory, removable memory device(s), communications capability (whether wireless or cable) with a remote monitoring station and a means of machine-to-machine internet communication. Conveniently, the control unit manages the operation of the apparatus by means of, for example, a signal bus that links generic slave modules that are each capable of controlling the behaviour of motors or actuators or pumps whilst interrogating feedback sensors.

Embodiments of the invention are typically arranged to measure force, electrical impedance, mechanical distension and/or mechanical displacement.

For each of these different options the preferred apparatus is arranged to perform analysis of the output signals generated by the sensor head(s), the analysis may include comparing at least one characteristic of the output signal against corresponding reference data in order to determine whether or not the output signal is indicative of a leak. The analysis typically includes an evaluation of said at least one characteristic over time. The analysis is typically dependent on the level of vacuum created in the vacuum chamber. For example, said reference data may be dependent on the level of vacuum. Said at least one characteristic of the output signal conveniently includes (and may consist of) the amplitude and/or the rate of change of the output signal and/or the rate of change of the rate of change of the output signal. Hence said analysis may comprise an evaluation of the amplitude of the output signal over time and/or the rate of change of the output signal, and preferably also against the level of vacuum in the chamber. Said reference data may comprise one or more corresponding reference output signals representative of a non- leaking package, or other item. The reference data may be obtained empirically.

As a package is subjected to an external vacuum it responds mechanically over a period of time rather than instantaneously. For example, the packaging may have some elasticity and may therefore continue to expand for a period of time until it reaches its elastic limit. Similarly, depending on the type of product inside the packaging, it may be that air (or other gas) is trapped within the packaging and that the secondary vacuum which is introduced within the packaging as a consequence of the external vacuum will cause outgassing of the product within the packaging over a period of time. These processes (and possibly others) mean that the mechanical measurements obtained from whichever sensor(s) are employed will not reach a stable value until after this period of time has elapsed. This time period may be several minutes and would delay a production line unacceptably. Preferred embodiments of the invention allow the time profile of the sensor output to be characterised to the applied vacuum in order to predict the future sensor output on the basis of its initial behaviour.

To this end, the apparatus is advantageously arranged to extrapolate a portion of an output signal received from said sensor head to create extrapolation data for use in said analysis of the output signal. In this context, extrapolation involves analysing the output signal in order to predict its future characteristics, and the extrapolation data comprises data for use in predicting future characteristics. The extrapolation may be based on any suitable mathematical model. The extrapolation analysis may be performed in real-time as the output signal is being received from the sensor head. This allows the apparatus to reach a conclusion about leak determination in a shorter period of time than would be possible by waiting for a sufficient portion of the actual output signal to be received.

Preferably, the extrapolation analysis is performed by analysing the gradient of the output signal of the sensor over time. An example of a suitable technique is given below: a) obtain the first derivative, i.e. gradient, of the time graph of the sensor output; b) track this first derivative signal until a peak is detected;

c) continue to track this first derivative signal until either time T has elapsed or the derivative signal has fallen by P1 percentage of the peak or has fallen to A1 absolute value (or by an absolute value); d) continue to track this first derivative signal until either time T has elapsed or the derivative signal has fallen by P2 percentage of the peak or has fallen to A2 absolute value (or by an absolute value); e) if time T has elapsed before either of these conditions has been reached then conclude that no leak is found; f) if the time of occurrence of a fall by P1 or to A1 is less than time T1 from the start of vacuum then conclude that a leak is found; and optionally g) if the interval between a fall by P1 and P2 or to A1 and A2 is less than period T2 then confirm leak found. The values for T, T1 , T2, A1 , A2, P1 and P2 may vary depending on the item under test and can be determined empirically prior to testing. In this example, it is assumed that P1 < P2, A1 < A2 and T1 < T2.

A second aspect of the invention provides an apparatus for detecting a leak, the apparatus comprising a vacuum chamber, means for generating a vacuum within the vacuum chamber and a sensor head located within the vacuum chamber, wherein said sensor head includes a plethysmographic sensing device.

A third aspect of the invention provides an apparatus for detecting a leak, the apparatus comprising a vacuum chamber, means for generating a vacuum within the vacuum chamber and a sensor head located within the vacuum chamber, wherein said sensor head is configured to produce an output signal whose characteristics are determined by interaction between said at least one sensing device and said item during a leak test, and wherein the apparatus is arranged to monitor at least one characteristic of said output signal over one or more time periods, and to make a leakage determination in respect of said item based on detected changes in said at least one characteristic over said one or more time periods. The apparatus may be said to extrapolate a portion of an output signal received from said sensor head to create extrapolation data for use in analysis of the output signal.

A fourth aspect of the invention provides an apparatus for detecting a leak, the apparatus comprising a vacuum chamber, means for generating a vacuum within the vacuum chamber and a sensor head located within the vacuum chamber, the sensor head includes means for contacting a package, or other target item, in said chamber said contacting means defining a perimeter around said sensing device such that, in use, engagement of said contacting means with said package, or other item, defines a portion of said package, or other item, with which said sensing device is operable.

A fifth aspect of the invention provides a method for detecting a leak in an apparatus comprising a vacuum chamber, means for generating a vacuum within the vacuum chamber and a sensor head located within the vacuum chamber, said method comprising using a plethysmographic sensing device to detect changes in one or more characteristics of a package, or other item, in said vacuum chamber in response to changes in the level of vacuum in the chamber.

A sixth aspect of the invention provides a method for detecting a leak in an apparatus comprising a vacuum chamber, means for generating a vacuum within the vacuum chamber and a sensor head located within the vacuum chamber, said method comprising producing an output signal whose characteristics are determined by interaction between said at least one sensing device and said item during a leak test, monitoring at least one characteristic of said output signal over one or more time periods, and making a leakage determination in respect of said item based on detected changes in said at least one characteristic over said one or more time periods. This method may involve extrapolating a portion of an output signal received from said sensor head to create extrapolation data for use in analysis of the output signal. A seventh aspect of the invention provides a method for detecting a leak in an apparatus comprising a vacuum chamber, means for generating a vacuum within the vacuum chamber and a sensor head located within the vacuum chamber, said method comprising causing a perimeter to be defined around said sensing device when in engagement with a package, or other item, in the chamber to create a portion of said package, or other item, with which said sensing device is operable.

Another aspect of the present invention provides an apparatus for detecting a leak, the apparatus comprising a vacuum chamber, means for generating a vacuum within the vacuum chamber and a sensor head located within the vacuum chamber and comprising at least one sensing device, wherein said sensor head is removably coupled to the apparatus.

Another aspect of the present invention provides an apparatus for detecting a leak, the apparatus comprising a vacuum chamber, means for generating a vacuum within the vacuum chamber and a sensor head located within the vacuum chamber and comprising at least one sensing device, wherein said sensor head is carried by an extendible device that is movable within said chamber to accommodate different sizes of items to be tested.

An eighth aspect of the invention provides a vacuum system, especially for a leak detection apparatus, the vacuum system comprising a vacuum chamber and vacuum generating means coupled to said vacuum chamber for creating a vacuum therein, wherein a fluid reservoir is provided between said vacuum chamber and said vacuum generating means, said vacuum generating means being coupled to said vacuum chamber for creating a vacuum therein, and at least one valve provided between the fluid reservoir and the vacuum chamber, said at least one valve being configurable to adopt a first state in which it at least partially restricts fluid flow from said vacuum chamber to said reservoir to the extent that said vacuum generating means is able to create a vacuum in said fluid reservoir, and a second state in which said at least one valve allows fluid flow from said vacuum chamber to said reservoir to the extent that said vacuum chamber can be evacuated by the action of said vacuum created in said reservoir. In its first state, said at least one valve may be closed to prevent, or substantially prevent, fluid from flowing from said vacuum chamber to said reservoir. Alternatively, said at least one valve may adopt a relatively restricting configuration in which it allows a reduced fluid flow from said chamber to said reservoir in comparison to the fluid flow allowed by said at least one valve in said second state. In said second state, said at least one valve may be open, for example so as to not restrict fluid flow from the chamber to the reservoir.

Said at least one valve may be configurable to adopt a third state in which it restricts fluid flow from said chamber to said reservoir to the extent that said vacuum generating means is able to maintain a desired vacuum level in said chamber, although this functionality may alternatively be provided by said at least one valve in its first state.

In any one or more of said valve states where said at least one valve is not closed, the preferred arrangement is such that said vacuum generating means is able to draw fluid from said chamber via said at least one valve, and typically via said reservoir.

Preferably, at least one bleed valve is coupled to said vacuum chamber arranged to act against said vacuum generating means, a desired vacuum level in said vacuum chamber being maintainable in use by balancing the action of said at least one bleed valve against the action of said vacuum generating means.

The preferred system is operable in a first mode in which said at least one valve adopts its first state to allow said vacuum generating means to create a vacuum in said reservoir. In said first mode, said vacuum chamber is not held under vacuum, for example so as to allow items to be tested to be placed in or removed from the chamber. This may be achieved by configuring one or more valves to establish fluid communication between said chamber and a gas source, conveniently the external atmosphere. The preferred system is operable in a second mode of operation in which said at least one valve adopts its second state to allow said vacuum in said reservoir to create a vacuum in said reservoir by drawing fluid/gas from said chamber. To this end, said chamber is preferably substantially isolated from other fluid or gas sources, e.g. by appropriate setting of said one or more valves, although one or more bleed valves may be open.

The preferred system is operable in a third mode of operation in which said at least one valve adopts said first state (where in said first state said at least one valve restricts but does not prevent fluid flow from said chamber to said reservoir) or said third state to allow said vacuum generating means to maintain a desired vacuum level in said chamber via said at least one valve. Advantageously, in said third mode said vacuum generating means is also arranged to create a vacuum in said reservoir. The third mode of operation may be effected while an item(s) is being tested in said chamber.

The invention also provides corresponding methods of creating a vacuum in the vacuum chamber.

A further aspect of the invention provides a leak detection apparatus comprising a vacuum chamber, vacuum generating means coupled to said vacuum chamber and arranged to generate a vacuum within the vacuum chamber, and a sensor head located within the vacuum chamber and comprising at least one sensor device, the sensor head being associated with constraining means configured to define a perimeter at least partially around said at least one sensing device, wherein, during leak testing of an item located in said vacuum chamber, said constraining means is arranged to constrain said item to define a portion of said item within said perimeter with which said at least one sensing device is co-operable

The preferred vacuum system allows relatively quick evacuation of the vacuum chamber as a result of the vacuum established in the reservoir. Moreover, the vacuum in the reservoir can be created while other tasks are being performed, e.g. during product testing or product loading/unloading.

In preferred embodiments, the leak detection apparatus is arranged for detecting a leak in a sealed packaged product. The sensor head is conveniently mounted on the roof of the vacuum chamber, preferably by means of said extendible arm. Preferably, the vacuum generating means has a vacuum pump coupled to the vacuum chamber via a vacuum line. Typically, a vacuum valve is provided in the vacuum line. Preferably, the vacuum valve is arranged intermediate the pump and the vacuum chamber. Ideally, a vacuum regulator is provided in the vacuum line. Preferably, the vacuum regulator is located intermediate the vacuum valve and the vacuum chamber.

Ideally, the sensor head is coupled to an electronic control unit. Advantageously, data relating to leaks and their frequency is documented and used for process control.

Faults or changes in the packaging material seal strength can be determined by comparison with normal, or reference, values stored electronically. Advantageously, the leak detection apparatus is capable of self diagnosing as it records the information from the, or each, sensor head. If the same sensor head gives irregular readings, this indicates a faulty sensor head. Preferably, more than one sensor head is provided in the vacuum chamber. Ideally, the sensor heads are spaced apart along the vacuum chamber, for example in a linear array. Preferably, the sensor heads are spaced along the roof of the vacuum chamber.

Preferably, the vacuum chamber has an opening for receiving one or more sealed packaged products or other items to be tested.

Ideally, the apparatus for detecting a leak includes, or is associated with, means for supporting at least one sealed packaged product, or other item to be tested.

Preferably, the supporting means comprises a conveyor, for example comprising a conveyor belt and drive means. Advantageously, the apparatus is incorporated into a production line to allow testing of packaged products coming from a flow wrap machine.

The preferred leak detection apparatus reduces downtime and rework and increases process yield as leaking packaged products found during the leak testing can be repackaged. Preferably, the conveyor has flights that traverse the direction of travel of the conveyor. Advantageously, the flights divide the conveyor into compartments, each of which receives, in use, a sealed packaged product, or other item to be tested.

Ideally, means are provided to cause relative movement between the vacuum chamber and the conveyor. Preferably, the means comprises mechanical components coupled to the vacuum chamber for moving the vacuum chamber towards and away from the conveyor. Ideally, the mechanical components are powered hydraulically, electrically and/or pneumatically. In an alternative embodiment, means are provided for moving the conveyor toward and away from the vacuum chamber. Ideally, the opening of the vacuum chamber has sealing means for sealing engagement with the conveyor, or other support on which packages, or other items to be tested, rest during use. Preferably, the sealing means forms a substantially airtight seal between the vacuum chamber and the conveyor enclosing the sealed packaged product therein.

For example, the sealing means may comprise a rubber seal (or a seal formed from other deformable material) provided along the edge of the opening. By way of example, the vacuum chamber may comprise a parallelepiped container, or other housing, having no bottom thus creating the opening.

Preferably, the one or more conveyors are coupled to the same electronic control unit which the at least one sensor head is coupled to. Advantageously, if a sensor head fails, then the leak detection apparatus is capable of running as normal with the electronic control unit operating the conveyors to leave an empty compartment on the leak detection conveyor under the faulty sensor head.

A further aspect of the invention provides a method of detecting a leak in a sealed packaged product, or other sealed item, the method comprising placing the sealed packaged product, or other sealed item, in a vacuum chamber, creating a vacuum in the chamber, and causing the packaging, or other item, to be in contact with a sensor head as the pressure in the vacuum chamber is changed, especially decreased. Typically, the method comprises measuring the force applied by the packaging to the sensor head, or measuring one or more other parameters, for example electrical impedance,

displacement and/or distension. Preferably, the method comprises analysing the measured data against corresponding reference data, for example reference data measured during or derived from a test on a non leaking packaged product, or other item.

Further preferred features of the invention are recited in the dependent claims, while still further advantageous aspects of the invention will become apparent to those ordinarily skilled in the art upon review of the following description of preferred embodiments and with reference to the accompanying drawings.

Brief Description of the Drawings The invention will now be described by way of example and with reference to the accompanying drawings in which like numerals are used to denote like parts and in which: Figure 1 is a schematic drawing of a first embodiment of apparatus for detecting a leak;

Figure 2 is a second schematic drawing of the first embodiment of apparatus for detecting a leak;

Figure 3 is a graph of force vs. time for a sensor head comprising a load cell;

Figure 4 is third schematic drawing of the first embodiment of apparatus for detecting a leak;

Figure 5 is a graph of force vs. time for a sensor head comprising a load cell;

Figure 6 is a fourth schematic drawing of the first embodiment of apparatus for detecting a leak;

Figure 7 is a graph of force vs. time for a sensor head comprising a load cell;

Figure 8 is a partial perspective cutaway view of a second embodiment of leak detection apparatus;

Figure 9 is a partial perspective cutaway view of a third embodiment of leak detection apparatus; Figure 10 is a perspective view of a fourth embodiment of leak detection apparatus;

Figure 1 1 A shows an embodiment of an extendible device for carrying the sensor head, the device being shown in a first position; Figure 1 1 B shows the device of Figure 1 1 B in a second position;

Figures 12A and 12B are schematic views of a vacuum system suitable for use with leak detection apparatus embodying the invention; Figure 13 is a graph showing the vacuum pulled by the vacuum system of Figures 12A and 12B; Figures 14A to 14C are schematic views of a preferred vacuum system suitable for use with leak detection apparatus embodying the invention;

Figure 15 is a graph showing the vacuum pulled by the vacuum system of Figures 14A to 14C;

Figure 16 is a schematic view of a first embodiment of a constraining device for use with embodiments of the invention; and Figure 17 is a schematic view of an alternative embodiment of said constraining device.

Detailed Description of the Drawings

Referring generally to the drawings there is shown an embodiment of an apparatus for detecting a leak indicated generally by the reference numeral 1 , the apparatus 1 comprising a vacuum chamber 2, means for generating a vacuum within the vacuum chamber 2 and a sensor head 3 disposed within the vacuum chamber 2. The apparatus 1 is particularly suited to detecting a leak in a sealed packaged product 4. The sensor head 3 includes a sensing device for detecting changes in at least one physical characteristic of the package 4 in response to pressure changes within the chamber 2. The sensing device is conveniently of the type that contacts the package 4 during use in order to detect changes. In preferred embodiments, the sensing device may be a plethysmographic sensing device (for example a sensing device arranged to detect changes in electrical impedance or mechanical distension); a force sensor (for example a load cell or a piezo-electric sensor) and/or a displacement sensing device, although other transducers may be used instead or in addition.

Advantageously, the sensor head 3 is removable from the apparatus to allow it to be interchanged with one or more other sensor heads (not shown). Each sensor head may comprise a different type of sensing device. This allows a user to select a sensor head 3 to suit a particular application. For example, a load cell may be well suited for use with a packaging having stiff walls and a membrane seal, e.g. a yoghurt pot, whereas a product wrapped in cling-film would be better tested by a plethysmographic sensing device. In Figures 1 and 2, the sensor head 3 is removably mountable on a fixed bulkhead within the chamber 2. Preferably, however, the sensor head 3 is removably mountable on an extendible device, e.g. an extendible arm, provided within the chamber. The arm is adjustable with respect to the chamber to allow the sensor head mounted thereon to be positioned with respect to the package 4. Typically the arm is arranged to support movement of the sensor head in a vertical direction (as viewed in the drawings) and/or in a horizontal direction. This allows the sensor head to be positioned to suit packages, or other items, of different sizes. Figures 1 1 A and 1 1 B show an example of a suitable extendible device 70 in the form of an extendible arm comprising a mechanical linkage 72 carrying a support or mount 74 for removably receiving a sensor head 75. The linkage 72 is carried by on a mounting block 76, which is mounted to the chamber. The linkage 72 is movable between a first, or raised, position (Figure 1 1 A) and a second, or lowered, position Figure 1 1 B to accommodate packages of different sizes.

In the embodiment of Figures 1 and 2, it is assumed by way of example only that the sensor head 3 comprises a load cell. The sensor head 3 is conveniently mounted on the roof 5 of the vacuum chamber 2. The means for generating a vacuum comprises a vacuum pump 7 (Figure 8) coupled to the vacuum chamber 2 via a vacuum line 9. A vacuum valve 1 1 is included in the vacuum line 9. The vacuum valve 1 1 is located intermediate the vacuum pump 7 and the vacuum chamber 2.

A vacuum regulator 12 is included in the vacuum line 9 intermediate the vacuum valve 1 1 and the vacuum chamber 2. The sensor head 3 is coupled to a control unit 14, which preferably comprises a suitable programmed microprocessor or other processor.

Advantageously, information about leaks and their frequency is documented and used for process control via the control unit 14. Faults or changes in the packaging material seal strength is determined by comparison with normal acceptable reference data, which may for example have been recorded during or derived from a test procedure using a non leaking packaged product. The reference data is conveniently stored on the control unit 14. Advantageously, the leak detection apparatus 1 is capable of self diagnosing as it records the information from each load cell/sensing device. Referring now to Figure 9, three (although alternatively there may be 2 or more) sensor heads 3 are spaced along the roof 5 of the vacuum chamber 2. The vacuum chamber 2 has an opening 16 for receiving, in this illustration, three sealed packaged products 4. The apparatus 1 also includes a conveyor arrangement comprising a conveyor belt 19 and associated drive means. Advantageously, the apparatus 1 , when incorporated into a production line, allows testing of packaged products 4 coming from, for example, a flow wrap machine (not shown). The leak detection apparatus 1 reduces downtime and rework and increases process yield as leaking packaged products 4 found during the leak testing can be repackaged.

In use, a sealed packaged product 4 is placed in the vacuum chamber 2 and the vacuum valve 1 1 is opened to reduce the pressure in the vacuum chamber 2 as a vacuum is created by the vacuum pump 7 via the vacuum line 9. It will be understood that references to "vacuum" made herein are intended to embrace "partial vacuum", as would be apparent to the skilled person. The vacuum created in the vacuum chamber 2 is limited to a pre-determined level to avoid over stressing the packaging material 51 and the seals, the pressure limitation being regulated by the vacuum regulator 12. In response to a reduction of ambient pressure in the chamber 2, the packaging 51 of the packaged product 4 swells, see Figures 4 and 6, to endeavour to keep the same pressure inside the packaging 51 as exists in the vacuum chamber 2. The arrangement is such that the vacuum level in the chamber 2 reaches a level beyond that required to fully inflate the packaging. When the packaging 51 reaches full inflation, the pressure inside the packaging 51 is equal to the pressure in the vacuum chamber 2. If the vacuum continues to increase beyond this point, the pressure in the packaging 51 stops falling (because the packaging cannot appreciably inflate any further, although a further increase of volume within the sealed packaging 51 may occur due to stretching of the packaging material (in cases where the material is stretchable)) and this tends to create a pressure differential between the inside of the packaging 51 and the vacuum chamber 2 whereby the pressure inside the packaging is greater than the pressure outside the packaging. . The sensor head 3 in contact with the inflated sealed packaging 51 measures the effect that this differential pressure has on the packaging. In preferred embodiments the sensor head 3 achieves this by plethysmographic sensing techniques, for example measuring the electrical impedance of the packaging 51 and/or measuring distensions of the packaging.

In the embodiment of Figures 1 to 6, however, this is achieved by measuring the force exerted on the packaging by the load cell (or other force-sensing transducer).

As the pressure differential begins to increase, as shown in Figure 5, the packaging 51 pushes against the load cell. The force it exerts on the load cell is proportional to the pressure differential. If there is a leak, the air/gas can escape from the packaging 51 and as differential pressure begins to increase air (or other gas) escapes from the packaging into the vacuum chamber 2 and causes a reduced force, or in the case of large leaks, little or no force to be exerted on the load cell/sensor. As a result of the leak, the differential pressure does not increase with increasing vacuum in the vacuum chamber 2, or increases at a reduced rate when compared to a non-leaking package. Therefore, leaks can be detected at three stages of the process depending on how bad the leak is. Firstly, if the leak is large, the increase of lifting force (as the package inflates) will be slower than that of a non leaking packaged product 4, which means the gradient on a graph of pressure v time is not as steep as shown in Figure 5. Secondly, the peak force exerted on the load cell 4 is lower for a leaking packaged product 4. Thirdly in the case of slower, or lesser, leaks, the load exerted on the load cell 4 peaks at an expected level and then begins to fall off as shown in Figure 7. Similar effects can be observed by a sensor head that monitors other parameters, e.g. electrical impedance of the packaging 51 or distension of the packaging 51. Hence by analysing the output signal from the sensor head 3, the control unit 14 can determine if a leak is present by one or more of the above characteristics of the output signal, and may also make a determination as to the type of leak. This can conveniently be achieved by comparing the amplitude, especially the peak amplitude and/or the rate of change of amplitude, of the output signal against corresponding reference data. Advantageously, the apparatus is arranged to extrapolate a portion of an output signal received from said sensor head. In this context, extrapolation involves analysing the output signal in order to predict its future characteristics. In preferred embodiments, the apparatus is arranged to monitor the first derivative, and/or the second derivative, of the output signal with respect to time, and to characterise the output signal based on the determined derivative value(s) during an analysis period, wherein the characterisation involves determining whether the derivative value(s) are indicative of a leak or no leak. The derivative value(s) may therefore be said to comprise data for use in the analysis of the output signal, in particular the extrapolation of the output signal. The extrapolation analysis may be based on any suitable mathematical model. The analysis of the output signal may be performed in real-time as the output signal is being received from the sensor head. This allows the apparatus to reach a conclusion about leak determination in a shorter period of time than would be possible by waiting for a sufficient portion of the actual output signal to be received. Preferably, the extrapolation is performed by analysing the gradient (first derivative) of the output signal of the sensor over time, and/or the rate of change of the gradient (the second derivative). Typically, the output signal is sampled at any suitable time interval in order to perform the analysis. An example of a suitable technique is given below: a) obtain the first derivative, i.e. gradient or rate of change, of the sensor output over time; b) track the first derivative until a peak is detected;

c) continue to track this first derivative signal until either time T has elapsed or the derivative value has fallen by P1 percentage of the peak or has fallen to A1 absolute value;

d) continue to track this first derivative value until either time T has elapsed or the derivative signal has fallen by P2 percentage of the peak or has fallen to A2 absolute value;

e) if time T has elapsed before any of the conditions of c) or d) has been reached then conclude that no leak is found;

f ) if the time of occurrence of a fall by P1 or to A1 is less than time T1 from the start of vacuum then conclude that a leak is found; and/or

g) if the interval between a fall by P1 and P2 or to A1 and A2 is less than period T2 then confirm leak found. The values for T, T1 , T2, A1 , A2, P1 and P2 may vary depending on the item under test and can be determined empirically prior to testing. In this example, it is assumed that P1 < P2, A1 < A2 and T1 < T2. T and/or T1 may be measured from a designated start time that is deemed to represent the start of the application of vacuum to the package in the chamber. In practice, the start time may be the time at which the vacuum level in the chamber reaches a threshold level (which may or may not be the same as the set vacuum level referred to hereinafter). T2 may be measured from the start time, or from another point in time after the start time. Preferably, T2 is measured from the time at which the peak gradient is detected. A preferred method for analysing the output signal is given below: a) obtain the first derivative, i.e. gradient or rate of change, of the sensor output over time; b) track the gradient until time Tmax measured from the instant of the onset of vacuum (e.g. defined as the instant when the chamber vacuum level first exceeds a threshold value S);

c) at each sample time, if the value of the gradient is greater than the gradient at any previous sample time then record this time as Tpeak and this value of the gradient as Gpeak;

d) if at any time (e.g. any sample time) during the interval to time Tmax the gradient value is found to have fallen from its previous maximum Gpeak by P1 percentage of Gpeak, or the gradient value is found to have fallen from its previous maximum Gpeak by A1 absolute value then conclude that a leak has been found (and the leak test may be stopped); and/or

e) if at any time (e.g. any sample time), for which the period elapsed since time Tpeak is less than T, during this interval to time Tmax the gradient value is found to have fallen from its previous maximum Gpeak by P2 percentage of Gpeak or the gradient value is found to have fallen from its previous maximum Gpeak by A2 absolute value then conclude that a leak has been found (and the leak test may be stopped);

f) optionally, after time Tmax, if the interval from Tpeak to Tmax is less than T then continue to track the gradient until an interval T has elapsed after time Tmax but during this period do not update the value of Gpeak or the value of Tmax;

g) if at any time during this additional interval until T has elapsed the gradient value is found to have fallen from its previous maximum Gpeak by P2 percentage of Gpeak or the gradient value is found to have fallen from its previous maximum Gpeak by A2 absolute value then conclude that a leak has been found (and the leak test may be stopped); h) if interval T is reached without detecting a leak and interval Tmax is reached without detecting a leak then conclude that there is no leak in the packaging.

The values for T, Tmax, A1 , A2, P1 and P2 may vary depending on the item under test and can be determined empirically prior to testing. In this example, typically P2 < P1 , A2 < A1 and T < Tmax. Tmax may be measured from a designated start time that is deemed to represent the start of the application of vacuum to the package in the chamber. In practice, the start time may be the time at which the vacuum level in the chamber reaches a threshold level (which may or may not be the same as the set vacuum level referred to hereinafter). T may be measured from the start time, or from another point in time after the start time. Preferably, T is measured from the time at which the peak gradient is detected.

The optional steps described in (f) and (g) cover the scenario where Tmax is reached without ever achieving a peak in gradient. Alternatively, it may be concluded that if Tmax is reached without a detected peak in gradient then there is no leak.

Optionally, if Tmax is reached without achieving an absolute value A3 for sensor output then it may be concluded that there is a leak in the packaging. Referring now to Figures 9 and 10, the conveyor belt 19 has flights 21 that traverse the direction of travel of the conveyor belt 19. Advantageously, the flights 21 divide the conveyor belt 19 into compartments each of which receives one (or more) sealed packaged product 4. Means are provided to cause relative movement between the vacuum chamber 2 and the conveyor belt 19, typically comprising mechanical components coupled to the vacuum chamber 2 for moving the vacuum chamber 2 towards and away from the conveyor belt 19. The mechanical components may for example be powered hydraulically, electrically or pneumatically. In an alternative embodiment, means are provided for moving the conveyor belt 19 toward and away from the vacuum chamber 2.

The opening 16 of the vacuum chamber 2 has a sealing member for forming an airtight seal between the vacuum chamber 2 and the conveyor belt 19 (when the two are brought into contact) enclosing the sealed packaged product(s) 4 therein. The illustrated vacuum chamber 2 comprises a parallelepiped container 26 having its floor removed creating the opening 16, although more generally it comprises an open-bottomed casing. Referring especially to Figure 10, the apparatus for detecting a leak 1 may for example be incorporated into a production line which, for example, comprises an inlet conveyor 31 , two leak testing conveyors 32 and an outlet conveyor 33. A distribution conveyor 34 creates a pathway between the inlet conveyor 31 and the leak testing conveyors a 32. A plurality of fans 35 are provided along the inlet conveyor 31. Advantageously, the sealed packaged products 4 are supplied from an area where the packaging has been heat sealed. The fans 35 cool the molten seals to further reduce the risk of damage to the integrity of the seals prior to leak testing. The distribution conveyor 34 has respective movable guide devices 37 substantially in alignment with each of the leak testing conveyors 32 and a bypass conveyor 41 for directing packaged products 4 down the predetermined leak testing conveyor 32 and/or the bypass conveyor 41. A collection conveyor 39 is provided along the opposite end of the leak detection conveyors 32 to the distribution conveyor 34. The collection conveyor 39 creates a pathway between the leak detection conveyors 32, the bypass conveyor 41 and the outlet conveyor 33. The bypass conveyor 41 creates a pathway between the distribution conveyor 34, the collection conveyor 39 and the outlet conveyor 33. Advantageously, if the leak detection apparatus

1 slows down and a build up of packaged products 4 occurs on the inlet conveyor 31 then some of the packaged products 4 can be sent down the bypass conveyor 41. If for example, one of the main leak detection conveyors 32 is damaged, then the leak detection apparatus 1 can continue to run at 50% inspection with the remainder of the packaged products 4 going down the bypass conveyor 41. This ensures that the leak detection apparatus 1 does not interfere with the output of the production line. The collection conveyor 39 is associated with means for diverting a leaking packaged product 4 to a testing and resealing area. One or more of the conveyors 31 , 32, 33, 34. 39 and 41 are coupled to an electronic control unit 14. The conveyors 31 , 32, 33, 34, 39 and 41 may be coupled to the same electronic control unit 14 to which the sensor head 3 is coupled. In use, a method of detecting a leak in a sealed packaged product 4 comprises placing the sealed packaged product in the vacuum chamber 2, creating a vacuum in the chamber 2 and causing the packaging 51 to be in contact with a sensor head 3 as the pressure in the vacuum chamber 2 is decreased. The force applied by the packaging 51 to the sensor head 3 is measured and analysed to determine if a leak is present. In the event of a leak, the sealed packaged products 4 identified as having a leak may be redirected to a testing and resealing area.

The electronic control unit 14 may store information relating to each batch of packaged products 4 tested. During normal run conditions, the leak detection apparatus 1 can accumulate information in relation to the frequency of leaks. In a modification, the sensor heads are each provided with side guards. This is due to the fact that as each pack inflates it acts as bellows creating a large lifting force on the respective sensor head. Since it is required to measure the force due to differential pressure, the side guards are provided in order to exclude the force due to the bellows effect. The guards around the sensor head reacts the load away from the sensor head allowing it to measure a much smaller force.

In typical applications, it is desirable to perform leak tests as quickly as possible. One way to improve the speed at which each leak test can be performed is to reduce the time taken to establish a desired vacuum level in the vacuum chamber. The desired vacuum level may be referred to as the vacuum set point, and is the vacuum level at which leak tests are performed, or at least initiated. The leak detection apparatus includes, or is connected to, a vacuum system for creating a vacuum in the chamber 2. When creating a vacuum in the chamber 2, typically there is a transient period, during which the vacuum level is not at the desired level, followed by an equilibrium period where the vacuum level settles at, or substantially at, the set point.

Figures 12A and 12B illustrate a vacuum system 160 that is suitable for use with leak detection apparatus embodying the invention. The system 160, which is similar to the vacuum system described in relation to Figures 1 to 9, comprises a vacuum pump 107 connected to the vacuum chamber 2 by a vacuum line 109. A valve 1 1 1 is incorporated into the line 109 between the chamber 2 and the pump 107. The valve 1 1 1 is operable to selectably allow fluid communication between the pump 107 and the chamber 2 and so to allow the pump 107 to create a vacuum in the chamber 2. Conveniently, the valve 1 1 1 is operable to selectably allow fluid communication between the chamber 2 and a source of gas (not shown), which may conveniently be provided by the air of surrounding environment, via a fluid line 1 13. Hence, depending on the setting of the valve 1 1 1 , the system 160 is operable in a first mode (Figure 12A) in which the vacuum pump 107 is able to create a vacuum in the chamber 2, or a second mode (Figure 12B) in which air (or other gas) is allowed to flow into the chamber 2 to replace the vacuum. Typically, the second mode is performed after the first mode and allows a tested package to be removed from the chamber and replaced with the next package. Preferably, a second valve 1 15, referred to herein as a bleed valve, is connected between the chamber 2 and a source of gas (not shown but which may conveniently be provided by the air of the external environment) via a fluid line 1 17. The valve 1 15 is operable to allow gas into the chamber 2 in order to control the vacuum level in the chamber 2 in said first mode of operation. The valves 1 1 1 , 1 15 may be controlled by any suitable controller, e.g. control unit 14.

Figure 13 illustrates the vacuum level in chamber 2 over time in the first mode of operation. In a typical embodiment using a 160M ³ /Hr vacuum pump, the system 160 takes 3-4 seconds for the vacuum to reach the set point and stabilise so that leak test sampling can begin.

Referring now to Figures 14 and 15, an embodiment of a preferred vacuum system is described. The vacuum system 260 is suitable for use with leak detection apparatus embodying the invention and is similar to the system 160, like numerals being used to denote like parts. The vacuum system 260 includes a fluid reservoir 262 connected (by fluid line 264 in this example) to, or otherwise in fluid communication with, the vacuum pump 207 so that the vacuum pump 207 can act to create a vacuum in the reservoir 262. A control valve 266 is provided between the reservoir 262 and the vacuum chamber 2.

The control valve 266 is adapted and arranged to control the flow of fluid between the reservoir 262 and the chamber 2, and more particularly the flow of fluid from the chamber 2 to the reservoir 262. To this end, the control valve 266 is conveniently incorporated into fluid line 209 between the reservoir 262 and the chamber 2. The control valve 266 is advantageously configurable to adopt a first state, in which it acts as a flow restrictor whereby it at least restricts the flow of fluid from the chamber 2 to the reservoir 262 to the extent that the vacuum pump 207 is able to create a vacuum in the reservoir 262 (advantageously while still creating (or at least maintaining) a vacuum in the chamber 2 (balanced against the action of the bleed valve 215 in preferred embodiments)), or a second state, in which the valve 266 is open to allow fluid to be drawn from the chamber 2 by the action of the pump 207 and/or by the action of any vacuum that has been created in the reservoir 262, to the extent that the desired vacuum level can be created in the chamber 2. Optionally, the control valve 266 may be closed in said first state so as to prevent, or substantially prevent, fluid from flowing from the chamber 2 to the reservoir 262. It is preferred however that the valve 266 is configured to act as a flow restrictor in said first state so that the pump 207 can simultaneously draw a vacuum in both the chamber 2 and the reservoir 262.

Optionally, where the control valve 266 is closed in the first state, the control valve 266 may be operable to adopt a third state in which acts as a flow restrictor whereby it restricts the flow of fluid from the chamber 2 to the reservoir 262 to the extent that it allows the pump 207 to maintain a set vacuum level in the chamber 2 (balanced against the action of the bleed valve 215 in preferred embodiments). Preferably, however, in the first state the valve 266 is configured to restrict the flow of fluid from the chamber 2 to the reservoir 262 to the extent that it allows the pump 207 to maintain a set vacuum level in the chamber 2 (balanced against the action of the bleed valve 215 in preferred embodiments), in which case a separate third state is not required. The control valve 266 typically comprises a restrictor valve.

It will be apparent that the required operating specification of the valve 266, e.g. its effective bore, when acting as a flow restrictor is dependent on the physical

characteristics of one or more other components of the system, for example the pumping capacity of the pump 207, the bore of the fluid flow lines, the specifications of the valves 21 1 , 215, the dimensions of the of the vacuum chamber 2 and/or the dimensions of the reservoir 262. In typical embodiments, the control valve 266 is provided separately to the valve 21 1. For example, in the illustrated embodiment, valve 266 is in line with valve 21 1 , between valve 21 1 and the reservoir 262. Alternatively, the valve 266 may be located between valve 21 1 and the chamber 2. In alternative embodiments, a single valve could be used in place of the valves 21 1 , 266, having the functionality of both valves. For example, the control valve 266 could be configured to be operable in another state in which it establishes fluid communication between the chamber 2 and a source of gas, and if necessary, a further state in which the valve is closed. The valves 21 1 , 215, 266 may be controlled by any suitable controller, e.g. control unit 14. In a first mode of operation of the system 260, the pump 207 is activated to create a vacuum in the reservoir 262 during periods where normal atmospheric pressure, or at least a non-vacuum pressure, is established in the vacuum chamber 2 (which may be achieved by opening the valve 21 1 , and optionally the bleed valve 215, to the external environment as described above.) This corresponds with, for example, periods during which products to be tested are being loaded into or removed from the vacuum chamber 2. The pump 207 is able to create a vacuum in the reservoir 262 because the respective state of the valve 21 1 and/or the valve 266 prevents or at least restricts fluid flow into the reservoir. For example, in typical embodiments, in the first mode, the valve 21 1 closes to prevent fluid flow to the reservoir 262 (in which case the state of the control valve 266 is not important, although it is preferably in its first state). Alternatively or in addition, the control valve 266 may adopt a state in which it at least restricts or prevents fluid flow into the reservoir 262 (in which case the state of the valve 21 1 is not important). Therefore, while products are being loaded/unloaded, a level of vacuum is created in the reservoir 262 which can be used to evacuate the vacuum chamber 2 in subsequent operations of the system.

The first mode of operation is illustrated in Figure 14B, in which it is assumed by way of exam pie that the valves 21 1 , 215 are open to allow the cham ber 2 to settle at normal atmospheric, or other non-vacuum, pressure. It is noted that one or both of these valves could subsequently be closed while the system 260 is still in the first mode of operation, provided the control valve 266 is in a state in which it prevents or at least restricts fluid flow into the reservoir 262.

In a second mode of operation, the control valve 266 is configured to adopt its second state and the valve 21 1 is configured to allow fluid communication from the chamber 2 to the reservoir 262 (and is closed to the external environment). As a result, the vacuum previously established in the reservoir 262 acts via the open valve 266 to evacuate the vacuum chamber 2. This is illustrated by way of example in Figure 14C, where the valve 21 1 is configured to allow fluid to be drawn from the chamber 2 by the reservoir 262. The bleed valve 215 may be open but is preferably closed during this process. It is found that using the reservoir in this way allows the desired vacuum set point in the chamber 2 to be established more quickly than for the system 160 of Figure 12. Once the desired vacuum set point in chamber 2 has been established, or substantially established, it may be maintained by the action of the pump 207, preferably balanced against the action of the now open bleed valve 215, as described for system 160.

Preferably, however, the control valve 266 is configured to adopt its first state (if it is flow restricting rather than closed in the first state), or the third state (if it is closed in the first state), as a result of which a reduced proportion of the pump's capacity is required to maintain the vacuum set point in the chamber 2. Advantageously, the remaining capacity of the pump 207 may be used to create a vacuum in the reservoir 262 for subsequent use in the second mode of operation. This may be regarded as a third mode of operation of the system 260. The third mode is similar to the first mode in that the control valve 266 and/or valve 21 1 are set to allow the pump 207 to build a vacuum in the reservoir 262 while some other activity is taking place in respect of the vacuum chamber 2 (e.g. testing, or product loading/unloaded), as is determined by the setting of the other valves 21 1 , 215.

Figure 15 is a graph illustrating the time taken to establish a settled vacuum level in the chamber 2 using the preferred vacuum system 260. The desired vacuum equilibrium state is reached more quickly than is the case for system 160. Preferably, in transition from the second to the third mode of operation, the control valve 266 is caused to adopt its flow restricting state before the vacuum in the chamber 2 has reached the desired level. This helps to reduce the time taken to reach the desired vacuum equilibrium level in the chamber 2. The preferred system 260 allows a vacuum to be built up in the reservoir 262 while a product test is running (since only a proportion of the vacuum pump's capacity is required to maintain the vacuum in the chamber). As a result the required size of the vacuum pump is reduced in comparison with the system 160, reducing cost and energy consumption. Also a smaller bleed valve 215 can be used, reducing cost and the amount of noise generated, as the valve 215 has only to balance with the restricted vacuum as opposed to the capacity of the vacuum pump 207.

The test cycle time for system 260 is reduced in comparison with the system 160 as a result of the improved time for establishing the desired vacuum level in chamber 2. For example, in one embodiment, it was found that product testing could start approximately 1 second after the chamber 2 was closed, in comparison with 3-4 seconds for the system 160. Advantageously, the sensor head 3, 75 includes a constraining device for constraining a portion of the package, or other target item, during measurement. The constraining device is configure to defining a perimeter around, or at least partly around, the sensing device such that, in use, engagement of the constraining device with said package defines a portion of said package, or other item, with which the sensing device is operable. The constraining device may for example comprise one or more contact members, e.g. one or more annular contact members, such as a collar, and/or a plurality of contact members, e.g. blocks, (which may or may not be contiguous) arranged to define the perimeter. The defined perimeter may be annular in that extends entirely or substantially entirely around the sensing device. However, one or more gaps may be left in the perimeter to ensure that the portion of the package within the perimeter is susceptible to changes in the vacuum level in the vacuum chamber. Typically, the constraining device constrains the surface area of packaging, or other item, that is affected by the applied vacuum and co-operable with the sensing device in order to amplify the effect being measured by the sensor head. This allows the impact of the vacuum in the vacuum chamber to be localised to a smaller region inside the constraining perimeter. In a typical implementation an annular, or substantially annular, constraining device is provided for pressing against the package. Preferably, the arrangement is such that the constraining device holds the packaging material against the contents of the package. In some cases, this may allow the external vacuum only to impact on the portion of the packaging that falls inside the constraining device around the sensor. The constraining means is particularly suitable for use with a load cell.

Figure 16 illustrates an embodiment of a constraining device 80 included in a sensor head suitable for use with apparatus embodying the invention. The constraining device 80 is located around the sensing device 82, which may for example comprise a load cell. In the embodiment of Figure 16, the constraining device comprises a plurality of contact blocks 86 for engagement with the package 4 and define said perimeter around the sensor 82. By way of example, two contact blocks 86 may be provided, one on either side of the sensor 82. In this embodiment, the contact blocks 86 do not form a continuous perimeter around the sensor 82. Gaps 88 are present between the respective ends of the blocks 86 (only visible on one side in Figure 16). This ensures that the portion of the package 4 between the blocks 86 is exposed to changes in vacuum level in the chamber 2. Figure 17 shows an alternative embodiment of the constraining device 180. The constraining device 180 is located around the sensing device 182, which may for example comprise a load cell. In this embodiment, the constraining device comprises an annular contact member 186 in the form of a collar. The collar 186 does form a continuous perimeter around the sensor 180. However, in this case gaps 188 are present in the upper surface of the collar 186 to allow the portion of the package 4 within the collar 186 to be exposed to changes in vacuum level in the chamber 2. The sensor head, including the constraining device 80, 180, is carried by a support 84 (not shown in Figure 17). In preferred embodiments, the sensor head is removably mounted on the support. The constraining device 80, 180 may be removable with the sensor head, or may remain attached to the support 84 when the sensor is removed. As such the constraining device 80, 180 may be regarded as part of the sensor head, or part of the support. In either event, it is associated with the sensor head such that it defines the required perimeter around the sensor 82, 182.

Constraining devices may be arranged to exposed the constrained portion of the package to changes in vacuum level in the chamber 2 by any suitable means, for example by configuring the constraining device to define one or more apertures (gaps) or channels. These may be provided in any suitable location, e.g. in any surface of the constraining device that is exposed to the vacuum chamber during use, or in the contact surface(s) of the constraining device. The invention is not limited to the embodiments described herein which may be modified or varied without departing from the scope of the invention.