Today's manufacturers of pharmaceutical products have been undergoing a continuous process of automation aimed at increasing production volumes, raising quality and countering cost pressures. Even in a well-managed, thoroughly automated and mechanized facility, the possibility of the product becoming contaminated with metal at some point in the production process is very probable.
Frequently, metal particles enter the production process along with the primary products. No matter when or where metal contamination can or might take place, one absolutely essential requirement of good quality management is that no flawed product ever be allowed to reach the market. To prevent this from happening, metal detection and rejection systems are now essential in the process stream to reliably identify and separate out all metal-contaminated products including those in aluminum foil. (Fig 1.)
A Rationale for Using a Precision Metal Detector
A piece of metal in a finished product can cause serious injury to the person who consumes the contaminated piece. Therefore, consumer protection is of primary importance in this context. Consumer protection when analyzed a little more closely is related to and an important component of brand-name protection. Just suppose you happen to be reading a regional newspaper and stumble across the headline "Consumer gravely injured by chip of metal in a (name a pharmaceutical company's product)" and the product in question is one of yours (lawsuits not withstanding)! Without wishing to disparage the consumer's injury, your first worry, of course, would be how much damage this is going to cause your company's reputation, much less its bottom line.
Figure 1: A Rationale for Using a Precision Metal Detector.
Another factor in the equation is machinery outage caused by even a tiny piece of metal wandering around in the production process. The result is lost time and money due to a production stoppage or even worse a plant shutdown. Therefore, to make a point, without belaboring the damage to a company's image- and health hazards aside we can correctly assume that every manufacturer wants to produce high-quality products and keep their machinery functioning at their peak capabilities. Looking at the bigger picture and avoiding the nightmare scenario of a plant closing, finding the metal before it becomes a major problem should be the goal of every plant manager or supervisor.
As I already mentioned, if the plant is working under intense cost pressure, its production resources must be reliable and economically efficient. It is therefore prudent to install precise metal detection and rejection systems at various strategic points along the production line rather than risk a possible plant stoppage or shut down caused by the failure to detect the problem in the first case. Thus when metal is detected and eliminated from the process the cost of the detector becomes a non-issue in comparison to the peace of mind and bottom line of the manufacturer involved.
The Benefits of Metal Detection
The benefits of metal detection follow directly from the reasons for using metal detectors in the first place. By systematically exploiting the advantages of metal detection and rejection with state-of-the-art technology, a manufacturer can invest all its energy to refining its products and developing new ones. They won't have to worry about lawsuits for immaterial damages or about the company's image suffering because of negative headlines.
Modern metal detectors do much more than just sound the alarm on metal particles. They have also become a front line instrument of many manufacturer's quality management systems and SOP protocols. These advanced features include automatic check prompting and rejection monitoring that is now integrated into the production data as is quantity of pieces, number and time of metal-detected messages, detailed fault messages, beginning of recording and detection sensitivity settings. The accumulated data is then sent via an interface (provided with the metal detector by the manufacturer) to a central computer for evaluation.
Basic Design of Metal Detectors
When sensitivity requirements are particularly high, metal detectors with a tunnel like opening for the product to pass through are the preferred configuration (Fig. 2).
The housing of the metal detector is made of stainless steel with deference to hygienic requirements, particularly in wet areas. The evaluating electronics unit is mounted either directly on the detector or at an operator's station situated a few meters/yards away. The measuring system itself is integrated into the detector and consists of three coils (Fig. 3).
Figure 2: The housing of the metal detector is made of stainless steel with deference to hygienic requirements, particularly in wet areas.
The coil at the center is the transmitter. Driven by an oscillator, it sets up a more or less high frequency alternating field. We will take a closer look at the effects of frequency later on
The transmitter is flanked by a pair of receiving coils situated equal distances away. The two receiving coils are wired up such that the voltages induced on either side by the transmitter cancel each other out. In the idle state the receiver voltage equates to zero and only changes if a piece of metal passes through the detector. The resulting signal is then amplified and processed electronically into a metal detected signal.
The size of the gate in a metal detector depends on the geometry of the product to be monitored: The smaller the gate, the better the base sensitivity.
Basically, there are two ways for metal to affect the magnetic field:
Figure 3: The measuring system itself is integrated into the detector and consists of three coils.
Ferromagnetic metals (those containing iron) and ferrites are most likely to cause concentration of the magnetic field lines (Fig. 4). Referred to as reactive, this effect is more or less pronounced, depending on the material's permeability, but will always be stronger than the other effect field displacement which we will go into next (see below).
At the point where the metal is located within the search tunnel, the field lines are subject to localized concentration. The result in homogeneity of the transmitting field produces different induced voltages in the two receiving coils, and the difference amounts to a metal-detected signal that can be evaluated by the detector.
Figure 4: Ferromagnetic metals (those containing iron) and ferrites are most likely to cause concentration of the magnetic field lines.
Due to the permeability of diamagnetic and paramagnetic metals (stainless steel and nonferrous metals), they prevent concentration of the magnetic field. On the contrary, the alternating field set up by the transmitter induces a voltage in the metal. Depending on its electrical conductivity, this produces a current, i.e., an eddy-current that in turn generates its own magnetic field. In accordance with Lenz's Law, the magnetic field appearing in the particle of metal opposes the exciting field. The two fields repel each other, and the field around the particle of metal is displaced, or distorted (Fig. 5).
The inhomogeneous state of the receiving field resulting from that phenomenon produces a measurable voltage in the receiving coils that can be evaluated as a metal detection signal.
Figure 5: The two fields repel each other, and the field around the particle of metal is displaced, or distorted.
Eddy-current formation pulls energy out of the excitation system so the effect is referred to as resistive.
In actual practice, both of these effects usually appear in combination. Depending on the transmitting frequency, a more or less pronounced eddy current will form on the surface of a ferromagnetic particle. This reduces the metal's permeability, perhaps to less than 1, depending on the size of the metal particle and on the working frequency. Outwardly, then, a piece of ferromagnetic metal acts as if it were diamagnetic.
Determining Factors of Metal Detection: Detectability of Various Metals as a Function of Transmitter Frequency - Basic Causality
As already mentioned, the effect responsible for the detection of metal changes with the frequency emitted by the metal detector and with the size of the metal fragment. Figure 6 illustrates this causality for spherical pieces of iron and of a nonferrous metal/stainless steel, each with a certain diameter. One product with a defined degree of conductivity (moisture) behaves like stainless steel (VA), but we will discuss that later on. As the curves clearly show, there is a certain optimal working frequency for the detection of various types of metal in combination with the product's own characteristics.
Figure 6: spherical pieces of iron and of a nonferrous metal/stainless steel, each with a certain diameter.
Phase Relations of Various Metals with Respect to the Transmitter Phase Angle
Each different kind of metal generates its own typical metal-detected signal specific to the transmitter voltage. It is useful to consider the metal-detected signal as a vector, with the phase angle representing the species of metal and the amplitude as the size of the particle.
Figure 7 illustrates the connection, i.e., it is a vectorial version of Figure 6 for a certain frequency.
Figure 7: the connection, i.e., it is a vectorial version of Figure 6 for a certain frequency.
The phase angles are subject to the same physical action mechanisms as those described before (3), and, as also already mentioned, the phase angle is function of the size of the metal fragment.
To begin with, the various species of metal iron (Fe), nonferrous metals (aluminum, Al; brass, br; bronze, bz) and nonmagnetic stainless steel (VA, e.g., 1.4301, 1.4401) have distinctly differentiable phase angles. Magnetic varieties of stainless steel, e.g., 1.4043, behave like iron, which is detected by way of its pronounced ferromagnetic effect.
The Product Effect
A product is said to have a product effect if it is electrically conductive, i.e., if it would commonly be referred to as somewhat moist. From the technical standpoint of metal detection, this kind of conductive product behaves like stainless steel or a nonferrous metal. Since its volume is significantly larger than that of a piece of metal, such a conductive product can also be expected to have a signal vector that is larger than that of the metal (Fig. 8). In order to detect and indicate a metal fragment in such a constellation, the detector must feature two-channel evaluation based on two different, independently adjustable sensitivity levels. Modern electronic equipment for metal detection should also be able to suppress the product effect. This suppression, or "learning", of the product effect is understood as the process of turning the product vector such that only the signal from the piece of metal to be detected is indicated in a given axis of evaluation (Fig. 9).
Figure 8: Since its volume is significantly larger than that of a piece of metal, such a conductive product can also be expected to have a signal vector that is larger than that of the metal.
Since, as already mentioned, the various types of metal have different phase angles, it is difficult to identify each different metal species unless each evaluation axis has its own separate sensitivity adjustment - like some metal detectors have the capability to do. Any change in the product effect - normally a result of a change in its phase angle - will be "noticed" by the tracking function, so a counterbalance can be initiated and a loss of sensitivity avoided. Figure 8 - Vectorial addition of a product effect and a detected-metal signal (Fig. 9).
Monitoring Metal Detectors
State-of-the-art metal detectors are equipped for automatic monitoring of their electronic functions. This keeps the appliance in good working order, but its product-specific detection sensitivity defies electronic checking, because it is not possible to simulate a product with a piece of metal embedded in it. Consequently, the product-specific sensitivity check must be performed with the aid of a test piece.
Figure 9: This suppression, or "learning", of the product effect is understood as the process of turning the product vector such that only the signal from the piece of metal to be detected is indicated in a given axis of evaluation.
Test Pieces for Metal Detectors
As discussed above, the measurable vector sum (= resultant) is the compound of the product vector plus the metal vector. Consequently, for a product with a product effect, it is important that the sensitivity test be conducted on an appropriately prepared original product. The test piece should not be situated on top or below the product, but somewhere within the product.
If the metal-detector's threshold sensitivity is to be checked for different species of metal, the respective test pieces will have to be inserted into an appropriate number of different products. The test pieces must, of course, be selected such that their signals are always situated above the metal detector's threshold sensitivity setting. For reasons of operational reliability, that threshold in turn should be situated well above the product signal. The achievable sensitivity with the original product has been referred to as the operational sensitivity. By contrast, the base sensitivity describes an instrument's basic detection sensitivity. For products with no product effect, the operational sensitivity is equal to the base sensitivity.
Gate Sensitivity Profile
The sensitivity of a metal detector is not equal and constant across the entire tunnel-shaped opening, or gate. This is because of the field distribution within the gate. The field density declines in inverse proportion to the distance from the point of origin. Consequently, the least sensitive point in the cross section of the gate is situated at the center (Fig. 10), so the test piece should always be made to pass through the center of the search-coil gate.
Detectability of Wires as a Function of Their Gate-Passing Orientation
As already described previously, the metal detector's magnetic field is generated by a flow of electricity through its transmitter coil. The resultant magnetic field has a certain distribution and orientation with respect to the gate. Regarding the detectability of metals, the extent of field-symmetry disruption is the decisive factor. Differentiation must be made between the reactive effect of field concentration and the resistive effect of field displacement resulting from eddy-current formation. The transmitter's magnetic field is only able to induce voltage in particles of metal that do not have the same orientation as that of the magnetic field lines. The induced voltage is at its maximum, when the piece of metal comes through at right angles (90°) to the field plane. Thus, the instrument's sensitivity for wires is directional, and, of course, it is also dependent on the type of metal to be detected (Fig. 11 and Table 1).
Figure 10: the least sensitive point in the cross section of the gate is situated at the center, so the test piece should always be made to pass through the center of the search-coil gate.
Correlation Between Iron Pellets and Wires With Least-Favorable Orientation for Detection
For a medium working frequency, Table 2 illustrates the relationship between the detectability of an iron pellet and that of various wires. It is postulated that all of the wires are of homogeneous structure. The comparison is much more difficult for the kind of stainless-steel chips that are likely to be encountered in actual practice. With its inhomogeneous structure, the stainless-steel chip reacts even more weakly than a wire with the transmitter's magnetic field. In the interest of test-data reproducibility, it is therefore not advisable to make test pieces out of stainless-steel chips, since it is practically impossible to obtain two test pieces that show identical behavior within the magnetic field.
Figure 11: The instrument's sensitivity for wires is directional, and, of course, it is also dependent on the type of metal to be detected.
With a view to ridding the test of orientational dependences, pellets of a certain size and material species can be used as test pieces. Test pieces of certified conformity can be obtained from some manufacturers such as Sartorius Mechatronics Corp.
Determining Factors for Operational Sensitivity
As may be deduced from what has been said up to this point, metal-detection technology is a very complex subject, and the ultimate results of detection are determined by many different factors. The following list does not purport to completeness, and the order of mention is no indication of relative importance:
1. Size of gate
2. Product conductivity (product effect)
3. Mode of conveyance: continuous/ intermittent
4. Working frequency of search coil
5. Metal species
6. Shape of metal fragment
7. Limiting ambient conditions
Rule of Thumb for Approximating the Sensitivity for a Dry Product
The sensitivity threshold for a dry product is estimable. The following rule of thumb applies to rectangular gates and a working frequency of roughly 150 kHz:
Fe ø [mm] = height of gate [mm] 250 +0.5 VA ø [mm] = height of gate [mm] m 250 +1.0
For each doubling of the gate width, the sensitivity drops by 10 .. 20 %.
Sensitivity for Products With Product Effect
The threshold sensitivity of the metal detector for products with a product effect can only be estimated on the basis of empirical data or, for better results, by way of an original-product test.
Metal Free Zone
The metal detector's stainless steel housing ensures that most of the magnetic flux lines remain within the housing. Nevertheless, a few lines of flux emerge out of the gate and close around the metal detector. This effect can be minimized by special casing-design measures, but it can never be eliminated completely. Detectors with reduced metal-free zones are now state of the art. Any lines of magnetic flux penetrating outward can also be influenced outside of the detector, of course, so it is customary to provide a metal-free zone directly in front of the detector's gate.
Separation and Removal of Detected Metal Fragments
It is one thing to detect a piece of metal, but quite another to remove it, though the latter is at least as important as the former. Depending on the type of product involved, and on how it is being handled, there are different approaches available for systematically detecting and separating out pieces of metal (Fig. 12). If the metal-contaminated product is coarse or lumpy, the alternatives include:
* Blow-off devices
* Swivel arms
* Tilting conveyors
* Telescopic ejectors
In most cases, blow-off devices, swivel arms and pushers are used in combination with:
* Belt conveyors or
* Plastic chain-link belts
For material in bulk, the best option would be a reversing belt, a diverting slide, or product marking in combination with belt stoppage. Pneumatically conveyed products are amenable to the use of Airtect ejectors (specially engineered for that purpose). The rejecter must ensure that the delivery pressure is maintained despite the ejection of a piece of metal. Separators suitable for liquid products include:
* Pinch valves
* Ball valves
* Three-way valves
* Diverter valves/flaps
Which separator is ultimately opted for depends on the product and the prevailing boundary conditions, e.g., cleaning in place, sterile-state service, no dead space, etc. Free-flowing/fluid products for freefall monitoring call for either:
* Diverter valve
* Diverter funnel.
In order to provide adequate protection for consumers as well as for one's own machinery, no production line should be without a metal-detection system comprising a metal detector and an ejector, possibly in combination with a conveyor belt.
The detection of miniscule particles of metal is an extremely complicated affair that depends on numerous different parameters. In many cases, the sensitivity needs to be determined with the aid of an original product. In choosing the right metal-detection system for the right product, the aid of an experienced producer of metal detecting equipment should be considered.