Red Guide

8/12/2019 Red Guide

1/33

THE GUIDE TOREDUCING METALCONTAMINATION

IN THE FOODPROCESSING

INDUSTRY

8/12/2019 Red Guide

2/33

Since its formation in 1988, Safeline has become the largest

metal detection manufacturer in the world with operations

in the U.K., U.S.A. and Europe. Design and manufacturing

experience from over 30 years, supports a worldwidedistribution network to advise on minimizing the risk of

metal contamination.

Safeline engineers provide additional customer support in

the form of training seminars in your facility, to discuss all

aspects of How to Develop an Effective Metal Detection

Program. If you would like more information about the

Safeline Microprocessor Metal Detector or to discuss a

seminar, please contact one of the Safeline offices.

THE METAL DETECTION GUIDE

8/12/2019 Red Guide

3/33

Copyright 1990 Safeline Ltd.

First Printing 1990

Second Printing 1994

Revised Printing 1996

THE GUIDE TO REDUCING METALCONTAMINATION IN THE

FOOD PROCESSING INDUSTRY

by

Andrew Lock

METAL DETECTION

8/12/2019 Red Guide

4/33


CONTENTS

Chapter Subject Page

1. Defining the Problem ...................................................................... 1

EQUIPMENT

2. Basic Principles ................................................................................ 2

3. Amplitude Detection - Zero Crossover............................................. 4

4. Sensitivity ........................................................................................ 6

5. Inspecting Wet or Conductive Products .......................................... 9

6. Conveyor and Reject Systems ......................................................... 11

7. Reject Timing.................................................................................. 16

MANAGEMENT OF QUALITY

8. Establishing An Effective Metal Detection Program ....................... 18

9. Reasons Why Your Program May Fail ............................................. 21

10. Developing a Foreign Material Control System .............................. 23

11. Comparing Metal Detector Brands ................................................. 26

12. Final Questions ............................................................................... 28

8/12/2019 Red Guide

5/33


MMMMMetal detectors are now accepted as essentialequipment by most food and pharmaceuticalprocessors. Many companies lay down strict

inspection standards in terms of detector sensitivity.

However, installing metal detectors will not necessarilyguarantee a metal-free product unless they form part

of an effective overall metal detection program.

This handbook has been written to help companies setup this program and prevent metal from gettingthrough.

The regulatory bodies in both the U.S.A. and U.K., aremaking increasingly strong recommendations for theuniversal inspection of all food and allied products bymetal detection equipment. For example, the UnitedStates Department of Agriculture (USDA) comments:

The extensive exposure of some products tometal equipment such as grinders, choppers,

mixers, shovels, etc., causes the possibility ofmetal contamination...therefore the use ofelectronic metal detectors is highlyrecommended....

(USDA Technical Services)

Revised guidelines were also issued by the Food andDrug Administration (FDA) to update the regulationsfor Good Manufacturing Practice and to establish newupdated, or more detailed provisions for the foodindustry to help ensure a safe and sanitary food supply.

Effective measures shall be taken to protectagainst the inclusion of metal or other

extraneous material in food. Compliance withthis requirement may be accomplished by using

sieves, traps, electronic metal detectors, orother suitable effective means.

(Federal Register Vol. 51 No 118)

1.1 Sources of Contamination

The sources of contamination are numerous and eventhe most stringent controls cannot prevent theoccasional incident. Working practices described inChapters 8, 9 and 10, will minimize the likelihood of

metal particles entering the production flow andmaximize the likelihood of reliably detecting andrejecting any that do.

Contamination normally comes from one of foursources:

Raw MaterialsRaw MaterialsRaw MaterialsRaw MaterialsRaw Materials

Typical examples include metal tags and lead shotin meat, wire and rust in wheat, screen wire in

powder material, tractor parts in vegetables, hooksin fish, staples and wire strapping from materialcontainers.

Personal EffectsPersonal EffectsPersonal EffectsPersonal EffectsPersonal Effects

Buttons, pens, jewelry, coins, keys, hair clips, thumbtacks, paper clips.

MaintenanceMaintenanceMaintenanceMaintenanceMaintenance

Screwdrivers and similar tools, welding slag andswar f fo ll ow ing repa ir s, copper wi re of fcut sfollowing electrical repairs, miscellaneous itemsresulting from inefficient cleanup or carelessnessand metal shavings from pipe repair.

In-Plant ProcessingIn-Plant ProcessingIn-Plant ProcessingIn-Plant ProcessingIn-Plant Processing

The danger exists every time the product is handledor passes through a process. Crushers, mixers,blenders, s l icers and transport systems a l l contribute. Examples include broken screens, metal

sl ive rs from mi ll ing mach ines and foi l from

reclaimed products.Identifying the likely source of contamination is animportant stage in developing an overall foreignmaterial reduction plan.

Inspecting raw materials will eliminate many large,easily detected pieces before being broken intonumerous, difficult to detect pieces.

1.2 Why Metal Detectors are Installed

To prevent damage to processing equipment.

To comply with stringent quality standards by majorcustomers.

These include high volume retailers, fast foodchains, food service and vendor certification

programs.

To avoid the cost and implications of consumercomplaints, adverse publicity, product recall andlitigation.

To win new markets and customers with high qualityproducts.

To comply with legislation such as Due Diligenceand FDA/USDA Directives.

1. DEFINING THE PROBLEM

8/12/2019 Red Guide

6/332


TTTTThe most common types of metallic contaminationin a broad range of industries include ferrous (iron),copper, aluminum, lead and various types of

stainless steel. Of these, ferrous metal is the easiest todetect and relatively simple detectors, or even magneticseparators, can perform this task well.

Stainless steel alloys are extensively used, particularlyin the food industry, and they are the most difficult todetect, especially the common non-magnetic gradessuch as 316 (EN58J) and 304L (EN58E). The non-ferrousmetals such as copper and lead fall between these twoextremes.

Only metal detectors using a balanced three-coil systemhave the capabilities to detect small particles of non-ferrous and stainless steel. The three coils are woundon a non-metallic frame or former, each exactly parallelwith the other. (Figure 1).

The center coil is connected to a high frequency radiotransmitter. The two coils each side of the center coilact as radio receivers or aerials. As these two coils areidentical and the same distance from the transmitter,they pick up the same signal and an identical voltage isinduced in each. When the coils are connected inopposition, they cancel out resulting in zero output.

When a particle of metal passes through the coilarrangement, the high frequency field is disturbedunder one coil, changing the voltage generated by afew microvolts. The state of perfect balance is lost andthe output no longer zero. The resulting signal isprocessed and amplified. It is this phenomenon which

is used to detect the presence of unwanted metal.

To prevent airborne electrical signals, or nearby metalitems and machinery disturbing the detector, thecomplete coil arrangement is mounted inside a metalcase with a hole in the center to allow the passage ofproduct.

Aluminum is normally used for the case but on some

applications, where frequent washdown is required,stainless steel may be used.

In addition to creating a screen, the metal case addsstrength and rigidity to the assembly. This is crucialfor satisfactory operation of the detector. Other specialmechanical and electrical techniques are essential toovercome many practical difficulties.

2.1 Mechanical Techniques

The metal case, itself, will have an effect on the stateof balance. In addition, microscopic movements of the

coils relative to each other, as small as 1 micron, cancause an out-of-balance voltage and a false detectsignal. One of the major design problems for metaldetector manufacturers is to design a totally rigid andstable system, unaffected by vibration from motors,pulleys, auto-reject devices, temperature changes,transportation and close-by machinery. The selectionof former material, coil specifications, and case design,are crucial. To increase mechanical rigidity further, somemanufacturers pot the detector with a material toprevent relative movement of the metal case to the coils.This helps produce a unit which is able to operate atmaximum sensitivity under normal factory conditions.

2. BASIC PRINCIPLES

8/12/2019 Red Guide

7/33


2.2 Electronic Techniques

Mechanical construction methods will minimize falsesignals from coil and case movements. However,temperature changes, buildup of product in theaperture, aging of electric components and slowchanges in the mechanical structure, will also contributeto an out-of-balance voltage. These can be eliminatedby various electronic techniques. Automatic Balance

Control continuously monitors this out of balancevoltage and automatically corrects it to zero. Thiseliminates the need for periodic fine tuning by anoperator and ensures the detector is permanently setto its optimum operating condition.

In a similar way, by controlling the frequency of theoscillator with great accuracy using quartz crystalcontrol, frequency drift is eliminated.

Automatic Balance Control and Quartz Crystal Controlwill not, themselves, enable the detector to detectsmaller pieces of metal. They will, however, enable thedetector to permanently maintain this sensitivity withno operator attention and without false reject signals.

For high performance over an extended period,Automatic Balance Control, Quartz Control and pottedheads are essential.

2.3 Ferrous-in-Foil Detection

When the product to be inspected is packaged inside analuminum foil pack or plate, a metal detector using thebalanced coil system cannot be used. An alternativedetector design is available which ignores the aluminumfoil but is sensitive to small pieces of ferrous contamination.Figure 2 shows the basic operating principal.

As a metal particle approaches the detector, it movesinto a powerful magnetic field which magnetizes theparticle. As this magnetized particle passes throughthe single coil, wound around the former, a smallvoltage is generated, which is subsequently amplified.

2.4 Metal Free Zone

Metal detectors need an area each side of the apertureto be free from metal structure, rollers and supports,to operate correctly. As a general rule, this needs to beapproximately 1.5x aperture height for fixed metalstructures and 2x aperture height for moving metal suchas reject devices or rollers. When space is restricted,such as with a short conveyor or for installation betweenscale and VFFS bagger, a patented Zero Metal FreeZone unit may be used.

8/12/2019 Red Guide

8/334


AAAAAs a metal particle passes through the detector,an output signal is generated which increasesto a maximum as it passes under the first coil,

falls to zero as it reaches the center coil, and increasesagain to a maximum as it passes under the third coil.

The signal will start to build up when the metal is some

distance from the coil. With a large metal piece, itcould be influencing the coil before it even arrives atthe detector. Figure 3 shows the signal generated by asmall and a large metal piece. This will be true for alltypes of detectors.

There are, however, two alternative methods ofinterpreting or processing this output signal, whichresult in different detector characteristics. One is knownas Amplitude Detection, the other is Narrow Zoneor Zero Crossover detection.

Amplitude DetectionAmplitude DetectionAmplitude DetectionAmplitude DetectionAmplitude Detection

When the signal from the metal particle exceeds apredetermined t rigger level, the detectoroperates. Figure 3 shows that a large metal piecebreaks the trigger level and is detected earlier than

a small metal piece.

Narrow Zone/Zero Crossover DetectionNarrow Zone/Zero Crossover DetectionNarrow Zone/Zero Crossover DetectionNarrow Zone/Zero Crossover DetectionNarrow Zone/Zero Crossover Detection

This technique gives a detect signal from the metal,when the signal changes polarity from a +ve to -ve or vice-versa. Figure 3 shows that this alwaysoccurs at the same point - under the center coil -independent of metal size.

There are certain advantages and disadvantages of eachmethod. The normal stated advantage of the narrowzone technique is that as the point of detection can be

accurately determined, independent of metal size, thevolume of rejected product can be minimized. Withamplitude detection, a large metal piece is detectedearlier and more good product may be rejected.

3.1 Multiple Metal PiecesThe major drawback of narrow zone is that thetechnique is not foolproof. In a typical production lineit is common for no contamination to occur for a longperiod and then several pieces to pass together, suchas when a sieve or grinder breaks up. If one metalpiece follows a second piece and the metals are of adifferent size, then the narrow zone detector can befooled and the smaller piece may not be detected.Figure 4 shows the signal from a small piece Afollowed by a larger piece B. The detector does notsee the two separate signals but the combined resultant

signal C. It can be seen that before this signal Chas a chance to change polarity and be detected, theeffects from the second piece swamps it. The first pieceis not detected. If a third large piece arrives, the firsttwo may not be detected and so on. This is a seriouslimitation of the narrow zone method.

3. AMPLITUDE DETECTION - ZERO CROSSOVER

8/12/2019 Red Guide

9/33


3.2 Effect of Outside Influences

The effects of vibration on the detector are also differentwith the two detection techniques.

Figure 5 shows that a signal from outside interference,such as vibration, occurs as a spike, with one polarityonly. Using amplitude detection, the trigger level maybe broken according to the level of vibration and a false

detect signal may occur rejecting good product. Withthe narrow zone technique, however, the signal doesnot change polarity and detection is inhibited. Twoinstances occur, however, where this is not effective.

If vibration occurs at the same time as a metal particle,the zero cross point can shift by up to 300mm (12).This would result in the reject device operating too earlyand the metal not being rejected.

If the outside interference comes from vibration on bothsides of the detector or from an intrusion on the metalfree zone of the detector, a dual spike will occur. Inthis instance, the zero cross occurs making the detector

overly sensitive to the interference signal.

8/12/2019 Red Guide

10/336


Metal TypeMagnetic

PermeabilityElectrical

ConductivityEase of

Dectection

Ferrous(Iron)

MagneticGood

ElectricalConductor

EasilyDetected

Non-ferrous(Copper, Lead)

Non-magneticGenerally

Goodor Excellent

RelativelyEasily

Detected

Stainless SteelVariousGrades

UsuallyNon-magnetic

UsuallyPoor

Conductors

RelativelyDifficult

to Detect

Fig 6

MMMMMany factors influence the operating sensitivityof a metal detector. Sensitiv ity figures needto be qualified with additional data, if they

are to be meaningful. It is fair to say that there iswidespread confusion and misinformation aboutsensitivity capabilities and specification.

The following factors effect sensitivity:

Type of metalShape of metalOrientation of metal

Aperture dimensionPosition of metal in the apertureEnvironmental conditionsProductOperating frequencyThroughput speed

When trying to determine an operating sensitivity orcompare capabilities of different detectors, thefollowing three factors are vital:

The sensitivity must be maintained permanentlywithout operator attention. An unstable unitrequiring constant attention is of no value.

The detector must not reject good product.

The detector must not give false reject signals fromvibration and outside influences.

4.1 Types of Metal

All metals fall into three main categories: Ferrous, Non-ferrous and Stainless Steel. The ease of detection will dependon their magnetic permeability (how easily they aremagnetized) and their electrical conductivity. (Figure 6).

Ferrous contamination is both magnetic and a goodelectrical conductor and is easily detected. Most metaldetectors are able to detect small ferrous particles. Non-ferrous metals such as copper, lead and aluminum, are

non-magnetic but are good electrical conductors and aregenerally quite easy to detect. Stainless steel comes in manydifferent grades, some magnetic and some austenitic(totally non-magnetic) and their conductivity is variable.For this reason, the grade of stainless steel should alwaysbe specified. In the food processing and pharmaceutical

industry 304L (EN58E) and 316 (EN58) are the two mostcommon grades. Poor sensitivity to these grades is a majorlimitation of many modern metal detectors. Wheninspecting wet or salty products, the problem of detectingstainless steel becomes even more acute.

A good indication of a detectors all round capability,is the sensitivity ratio between ferrous and the mostdifficult to detect grade of stainless steel 304L (EN58E).This ratio can be as good as 1:1.5 and as poor as 1:2.5.This has a major effect on the detectors ability to detectcontamination such as slivers and screen wire whichexhibit an orientation effect.

4.2 Shape of Metal

Metal spheres are used as a standard to determinedetector capabilities. There are two reasons for this.

Spheres are available in a range of metals anddiameters.

A sphere has a constant shape which ever way it ispresented to the detector. It has no orientationeffect.

The sensitivity of a detector is usually defined as thediameter of a metal sphere of a specific metal which isjust detectable in the center of the aperture.

4.3 Orientation Effect

If a non-spherical particle of metal such as swarf orwire passes through a metal detector, it will be easierto detect when passing in one orientation comparedto another. This is known as orientation effect. Figure7 shows that a piece of ferrous wire is in the mostdifficult orientation to detect when it is at 90 to thedirection of flow and the easiest when aligned along

the conveyor belt. Non-ferrous and stainless steel wiresare just the reverse. If this type of contamination islikely, care should be taken to ensure that the detectoris, in fact, capable of detecting it.

The orientation effect is only evident when the diameterof the wire is less than the spherical sensitivity of themetal detector.

With the detector sensitivity set at 1.5mm diameter, forexample, only wires thinner than 1.5mm diameter willshow the orientation effect. If the detector sensitivity isincreased to 1.0mm, only wires less than 1.0mm diameter

4. SENSITIVITY

8/12/2019 Red Guide

11/33


Fig 8

RELATING THE SPHERICAL SENSITIVITY OF A DETECTOR TO ITS ABILITY TO DETECT LONG THIN PIECES.

will cause a problem. If the diameter of the wire is onlyabout 1/3 the diameter of the detectable sphere, it maynot be detectable whatever its length.

Clearly, to minimize the orientation effect, it is better tooperate the detector at the highest possibly sensitivity.

This, however, may cause other problems. As sensitivitylevels increase, the problems of drift become more acuteand with some detectors, nuisance false rejects willincrease to an unacceptable level. The benefit of astable detector (See Chapter 11) becomes even moreimportant.

Figure 8 compares a detectors ability to detect four differentwire samples at various detector sensitivities. The left handcolumn shows four alternative sensitivities. As an example,when operating at 1.5mm diameter, the piece of copperwire would need to be 9mm long to guarantee detection.At a 2.0mm sensitivity, this would increase to 26mm. Itcan be seen that a small change in detector sensitivity willmake a great difference in its sensitivity to wire pieces.

The solution is to operate the detector at the highestpossible sensitivity. Auto Balance, Quartz Control andpotted head, will help this to be achieved.

Fig 7

8/12/2019 Red Guide

12/338


4.4 Aperture Dimensions

A large aperture is less sensitive than a small aperture.Both aperture width and aperture height have aninfluence on the detectors sensitivity, but changes inthe aperture height or smaller aperture dimension willhave the greater effect.

The geometric center is the least sensitive part of the

aperture, the corners are the most sensitive and all otherpoints will lie somewhere between. The differencebetween these two extremes is known as the sensitivitygradient and will depend on the design of the coil andformer assembly.

A large gradient may make the detector undulysensitivity to contamination in the conveyor belt orpacking material. A perfect detector would have nogradient and equal sensitivities at all points.

4.5 Environmental Conditions

Metal detectors are influenced to varying degrees byadverse environmental conditions such as airborneelectrical interference, plant vibration and temperaturefluctuations. These effects become even more acutewhen operating at high sensitivities. Ovens, freezingtunnels and hot water washdown, all create thermalshock which can result in false reject signals. UnlessAutomatic Balance Control is used, which eliminatesthe problem, the solution is usually for an operator toreduce the sensitivity of the machine. For this reason,when comparing detector capabilities, testing undercontrolled laboratory conditions is not too meaningful.Sensitivity specifications should be made for normal

production conditions with the detector stable. Somedetector manufacturers are more conservative thenothers when quoting sensitivities.

4.6 Product

Dry product such as confectionery and cereals, isrelatively easy to inspect and sensitivity charts can beused to calculate the expected operating sensitivities.When inspecting wet, conductive product such as freshmeat, the situation is different.

The wet product creates an interference signal in the

detector which needs to be canceled out beforeinspection can begin. This tends to reduce thesensitivity of the detector in a way that cannot becalculated empirically. To minimize the effect, a loweroperating frequency is often selected in the range 10-50 KHz. This reduces the interference signal from theproduct but also reduces the sensitiv ity of the detector,particularly to stainless steel.

To determine production line sensitivity, actual producttesting is essential and generally, the result shows aslight improvement in ferrous sensitivity but a reductionin non-ferrous and stainless steel sensitivity.

4.7 Inspection Speeds

Minimum and maximum inspection speeds are seldoma limiting factor for metal detectors, particularly onconveyor type applications.

The limit will vary from manufacturer to manufacturerbut will be a function of the detector aperture height.Typically this would be around 8M/sec. (1500 ft./min.)for a 125mm (5) high aperture. Minor modificationsare usually possible to extend this range further.

The limit of performance can be reached when

attempting to inspect on pneumatic pipelines at speedsin excess of 35M/sec. (6000 ft./min.).

More important than the absolute maximum andminimum, is a uniform sensitivity over the full speedrange. This is not universal to all detectors.

8/12/2019 Red Guide

13/33


PPPPProducts such as cheese, fresh meat, warm bread,

jam and pickles, can create a signal in a metaldetector even when completely free of metal. This

is known as product effect and is caused by the saltor acid content making the product electricallyconductive.

To make inspection possible it is necessary to eliminateor reduce this product signal. This can be done in oneof three ways:

Sensit ivity ReductionSensit ivity ReductionSensit ivity ReductionSensit ivity ReductionSensit ivity Reduction

By progressively reducing the sensitivity of the metaldetector, the signal from the product is made

smaller and smaller until it is no longer detectable.Despite the detector also becoming less sensitiveto all metals, it is usually the preferred option when

product signals are smal l.

Frequency ReductionFrequency ReductionFrequency ReductionFrequency ReductionFrequency Reduction

The operating frequency of a metal detector isgeneral ly in the range 10 KHz to 500 KHz. Byselecting a frequency towards the low end of thisrange, the signal from product effect becomes

smaller. Unfortunately, the signal from non-ferrousand stainless steel also becomes smaller, makingthe sensitivities to these metals lower.

Product CompensationProduct CompensationProduct CompensationProduct CompensationProduct Compensation

Special electronic circuits are available which canamplify and filter the signals from the detector by

differing amounts according to their characteristics.The filters are operator adjustable to take accountof a broad range of product signals. This techniqueis known as product compensation and generallyhas the effect of minimizing the product signal,improving the detectors sensitivity to ferrous metal,

reducing sensitivity to non-ferrous and stainlesssteel and, additionally, making the detector moreprone to vibration from motors, reject devices andother nearby machinery.

Inspecting conductive product is always a compromiseand, in practice, a metal detector manufacturer willuse a combination of all three techniques to give thebest operating performance.

The effects of vibration and drift from temperaturevariations are more pronounced on product effect lines.Automatic Balance Control, Quartz Crystal FrequencyControl and potting the detector head will help create

total stability. This overcomes a common problem oftenexperienced by users, of a gradual increase in theamount of rejected product which, when re-inspected,is found to contain no contamination.

The following section gives a more detailed explanationof product compensation.

5.1 Product Compensation - A Detailed Look

The signals created by various metals as they passthrough the coils of a metal detector can be split intotwo components, resistive and reactive, according to

5. INSPECTING WET OR CONDUCTIVE PRODUCTS

8/12/2019 Red Guide

14/3310


the conductivity and magnetic permeability of themetal. The signal from ferrous metal (iron) is primarilyreactive and stainless steel is primarily resistive. Figure9 shows in vector diagram form, the signals from anumber of different metals as they pass through thedetector. They show the signal increase to a maximumas they pass under the first coil, decay to zero as theypass under the center coil and again increase to a

maximum when passing under the third coil.The signals vary in amplitude (length) according tometal size and phase (direction) according to theresistive and reactive components which, in turn,depend on metal type.

The other points are evident from the diagram. Signalsfrom ferrous metal are larger than signals from the samesize piece of non-ferrous or stainless metal. Also, signalsfrom vibration and outside interference as always alongthe horizontal reactive axis.

To improve sensitivity of the metal detector to thestainless signal and reduce the sensitivity to vibration,special circuits can be used to amplify the signals bydiffering amounts, according to phase. This techniqueis known as Phase Sensitive Detection (P.S.D.). It isshown in diagrammatic form in Figure 10. The P.S.D.

is shown as a long thin oval called the detectionenvelope. For a signal to be detected, it must passoutside this envelope. Large signals from vibration arerequired before passing outside the envelope and beingdetected while only small signals from stainless steelare necessary. This is the most satisfactory operatingcondition.

A problem, however, occurs when inspecting aconductive product such as cheese. The large product

signal (Figure 11) passes outside the envelope and willbe detected each time, even when metal free.

By reducing the sensitivity of the detector, all the signalswill become smaller until the product signal no longerpasses outside the envelope. Inspection will then bepossible. For small product effect this will usually bethe preferable solution.

An alternative solution is shown in Figure 12. Thedetection envelope can be rotated electronically, until

it is aligned with the product signal. This is known asProduct Compensation and can be carried out by theuser. The product signal no longer passes outside theenvelope so normal inspection is again possible. Usingproduct compensation, however, has its drawbacks.Large signals from stainless steel are needed to passoutside the envelope so the detector becomes lesssensitive to these metals.

At the same time, small signals from vibration will nowpass outside the envelope and be detected. Undue

sensitivity to vibration is often the limiting factor wheninspecting with product compensation.

The exact phase of any product cannot be calculatedfrom data on salt content or on pH. For the samereason, detection sensitivities cannot be calculated.Product testing is essential to determine the detectorssensitivity to a range of metals. This service is usuallyavailable from metal detector manufacturers.

5.2 Automatic Product Compensation

Accurately adjusting the product compensation control

needs experience if optimum performance is required.If a number of different products or pack sizes are tobe checked on the same production line, adjusting thedetector for each new product can be time consuming.The recent introduction of microprocessor based metaldetectors has resulted in major improvements wheninspecting conductive products. By switching toLearn or Automatic Compensation mode, productcompensation can be set automatically, with nooperator involvement These settings can then beentered into a memory to allow immediate recall whenchanging product, or even allow remote setup of thedetector from a central computer.

8/12/2019 Red Guide

15/33


1

TTTTThe design of the conveyor system which transportsthe product through the detector, must meetcertain strict criteria if it is to avoid influencing

the detector in any way.

A metal detector conveyor is much more than a modifiedtransport conveyor. The design of both conveyor and auto

reject device will have a major impact on the effectivenessof your overall metal detection program.

Unless special precautions and design techniques areincorporated, eddy current loops and static buildup caninfluence the detector and downgrade sensitivity,causing interference and probable false rejections.

Metal detectors emit a high frequency radio signalwhich cause tiny eddy currents to flow all around themetal structure of the conveyor. These eddy currentshave no effect on the detector if they remain constant.However, if the conveyor structure has an intermittent

joint of variable resistance, even within a remote distancefrom the detector, the eddy currents briefly change whichcreates a large interference signal to the detector.

Typical sources of eddy current loops are any metal tometal contact such as a bolted conveyor assembly orsupports, pulley shafts and bearings, chain drives andguards, reject supports and metal conduit clamps.Frequently, oxidation of joints or changes in bearinglubrication will cause problems to increase with time.

6.1 Belting Types

A number of factors need to be considered in choosinga suitable belting material. Static charges can build upwith some, particularly when running over plastic skidplates and plastic coated rollers and pulleys. Specialantistatic belting can cause a problem as it is oftenmade with conductive carbon fillers or additives, whichwill adversely affect the detectors performance,particularly when the belt joint passes through theaperture. With any type of belt the joint must be metalfree and made in such a way to prevent product buildupor an accumulation of grease. A vulcanized or glued

jo int at 45 or an in terlocking finger jo int he lpsminimize this effect. Metal fasteners, or sewn and laced

joints, are unsuitable.The belt material itself must also be totally metal free.Tiny metal specks in the material are extremely difficultto find. Belt manufacturers producing consistently highquality, metal-free belting, would almost certainly needto use metal detection equipment to inspect their rawmaterials.

Flat, dished, ribbed, cleated and molded flexible wallbelt are all acceptable. Solid plastic chain belts of theIntralox style and round urethane belting running ingrooved rollers, are ideal where spillage is likely,requiring frequent washdown.

Endless double pass belts (Figure 13) offer a numberof advantages in many applications, including rapidreplacement. However, as the face of the belt passesover a roller they are not suitable for transporting wetor sticky product such as meat trimmings.

6.2 Product TransferTransfer onto the conveyor system needs specialconsideration when the end rollers are large or theproduct small. If the distance D between rollers is morethan half the product length, reliable transfer will notbe possible. Small non-powered intermediate rollersor a dead plate positioned between the two rollers,are usually effective. (Figure 14). Single or double knifeedges (Figure 15) permit transfer of very small itemswhere product registration has to be retained, such asrows of confectionery at the outlet of an enrober. Stickyproduct such as raw dough and meat and bulk loose

product such as loose peanuts, can be transferred bycascade (Figure 16).

6. CONVEYOR AND REJECT SYSTEMS

8/12/2019 Red Guide

16/3312


For jar inspection the detector system may be positionedalongside the existing transport conveyor and productguides divert the jars from the line onto the detectionsystem. Acceptable products are then diverted back.When contaminated items occur, the product guide canbe moved pneumatically to allow the item to be rejectedfrom the line. (Figure 17).

6.3 Transfer Speed

To allow easier identification of the contaminated items,it is often useful to accelerate the product through thedetector to create an increase in product spacing.

When packs are very close, the detector maybe unableto determine which is contaminated. Two or threepacks may need to be rejected to be sure of catchingthe right one. By increasing the detector conveyorspeed, product spacing is increased, permitting theindividual items to be identified.

When inspecting bulk and loose product, the burdenheight can be reduced by accelerating the product ontransfer. This has the advantage of minimizing the

volume of rejected product and permits a lower

detector aperture, resulting in higher sensitivities.(Figure 18).

6.4 Automatic Rejection Systems

Ineffective reject systems are probably the weakest linkin most detection systems and result in contaminationnot being effectively and reliably rejected from the line.

A correctly specified system should be foolproof andcapable of rejecting all contaminated product underall circumstances, independent of the frequency ofoccurrence or the location of the metal inside theproduct.

Occasionally auto rejection is not used. The operatoris expected to remove contaminated product when theconveyor is stopped on detection, or the sound of abell, or when a bright plastic disc is dropped on thebelt from an electric disc dropper designed to operateeach time metal is detected. All these solutions arehigh risk and depend entirely on the efficiency of the

line operator.The choice of the most appropriate reject system willdepend on a number of factors and the advice of thedetector manufacturer should always be sought.

Options available include:

Air BlastAir BlastAir BlastAir BlastAir B last - Ideal for light, single line product. (Figure19).

Punch/PusherPunch/PusherPunch/PusherPunch/PusherPunch/Pusher - Single line discrete, spaced andoriented product. (Figure 20).

8/12/2019 Red Guide

17/33


1

Sweep/Diverter ArmSweep/Diverter ArmSweep/Diverter ArmSweep/Diverter ArmSweep/Diverter Arm - Random and non-orientedproduct. (Figure 21).

End FlapEnd FlapEnd FlapEnd FlapEnd Flap - Bulk or discrete multiple items on wide belt.(Figure 22).

Retracting BandRetracting BandRetracting BandRetracting BandRetracting Band - End pulley retracts to create a gapin the line. Very reliable on most applications. (Figure23).

6.5 Inspection of Liquids

Inspection of pumped liquids and slurries can beachieved by replacing a short section of the stainlesssteel transport pipe by a food quality plastic pipe andpassing it through a metal detector (Figure 26). On

detection of metal, a sanitary three-way valve canoperate to divert the contamination or, alternatively,the pump can be stopped and the contaminationflushed out manually. Typical products suitable forpipeline inspection include liquid chocolate, ice cream,soup and meat slurry.

Reverse BeltReverse BeltReverse BeltReverse BeltReverse Belt - Ideal for bulk, random or sticky product.

(Figure 24 and Figure 25).

For product likely to solidify if pumping stops, such asliquid chocolate, the throughput pipe can incorporatea hot water jacket (electric wire wrap heating cannotbe passed through a detector). Hot water jackets alsoprevent an accumulation of fat on the pipe interiorwhen pumping certain types of sausage meatemulsions.

Special metal fittings may also be used to permit directconnection to sausage pumps or other systems. Tri-

Clamp, I Clamp, Sanitary, Vemag, Handtmannand similar connectors are generally available.

Pumped product is seldom totally homogeneous. Voidsand bubbles frequently occur and this can causeproblems when adjusting the detector for optimumperformance, especially for highly conductive products.

Under normal conditions, product is passing under bothcoils of the detector. Product effect tends to cancelout and the detector can be adjusted to give highsensitiv ity. If, however, a void or bubble appears as itpasses under the first coil, the detector will sense alarge product difference and a false reject will occur.

8/12/2019 Red Guide

18/3314


(Figures 27 and 28). It is possible to adjust the detectorto eliminate the product signal, but unless the bubblesoccur frequently, or at a predictable time (such as witha pump startup), it may take a long time. In theseinstances, automatic product compensation will be ofno help.

no moving parts such as motors, gears, rollers or beltsand with the relatively high volumes which can passthrough a smal l detector opening, very h ighsensitivity can be achieved.

The detector and auto reject should be mounted on arigid framework with sufficient space between themto ensure that metal contamination is always rejected.(Figure 29). Product flow should be continuous free

fall or batch free fall. The application is not suitablewhen product backs up the throughput pipe andmoves slowly.

Frequently the overall system height is a limitation tothe use of gravity feed systems - particularly where littleheadroom exists.

The fol lowing l imit ing var iables have a directrelationship on the overall system height.

Initial Fall Height of the ProductInitial Fall Height of the ProductInitial Fall Height of the ProductInitial Fall Height of the ProductInitial Fall Height of the Product

This will determine its velocity at the point of

inspection and also the time taken to arrive at thereject point.

The fall height should ideally be reduced to aminimum by locating the equipment as close as

possible to the point of initial fall.

Detector ApertureDetector ApertureDetector ApertureDetector ApertureDetector Aperture

This will determine the metal free zone of thedetector and, in turn, the height of the inputflange above the detector and the closest pointthe reject device can be located. Using the

patented Zero Metal Free Zone technology wi ll

keep this to a minimum. The aperture height willalso determine the distance the reject chute musttravel to reject product.

System Response TimeSystem Response TimeSystem Response TimeSystem Response TimeSystem Response Time

This covers the speed of response of the relay orsolid state output, air solenoid, air cylinder and thetime taken to move the reject flap to the reject

posit ion.

Reject AngleReject AngleReject AngleReject AngleReject Angle

The reject angle must not be too large to createblockage or bridging. As the length of the rejectflap is reduced, the reject angle increases. An angleof between 25 to 30 is considered a maximumfor most products.

Reject DesignReject DesignReject DesignReject DesignReject Design

Product buildup on the reject device, a drop in airpressure and aging of bearings, will slow down thespeed of response. A suff icient safety margin isneeded in the design to ensure metal is rejectedwith 100% accuracy.

6.6 Inspection of Powders and Granules

Any free flowing powder or granule such as peanuts,rice, plastic pellets, milk powder and cocoa beans,can be inspected under free fall conditions using agravity feed free fall detector and a high speeddiverter valve. Under normal conditions there are

8/12/2019 Red Guide

19/33


1

The following table shows the relationship betweenpipe diameter, sensitivity and typical throughput ratesfor gravity feed systems manufactured by Safeline.

pouch. Up to three fil lers - hence three detectorsmay be needed on a single pouch maker.

The foil pouches must be accurately indexed forfilling. A signal can be taken from the indexingmechanism to the shift register timer of the metaldetector and used to provide a synchronizedrejection of contaminated packs at some remote

point away from the fil ler.

VVVVVertical Form Fill Sealertical Form Fill Sealertical Form Fill Sealertical Form Fill Sealertical Form Fill Seal

A ZM FZ detector can be f itted between acomputerized scale and a vertical form fill seal bagger(Figure 30). Often a large detector of 150mm (6)or 200mm (8) is required, which would require alarger metal-free zone. The patented ZMFZtechnology allows high sensitivity levels to bemaintained, without false rejects, in the minimum

space, making instal lation easier and avoiding thedanger of product breakage. On detection of metal,the complete line may be stopped or the sealing

jaws inhibited, to double bag.

6.7 Special Applications

Installing a custom designed metal detector directlyonto, or into, a packaging or process machine, can havea number of distinct advantages for both the user and

the supplier of the original machinery. It is a trendthat is likely to continue. Where installation is requiredin a restricted space, the patented Zero Metal FreeZone (ZMFZ) detector can be used, which allows metalstructures and components to be positioned very closeto the detector without interference.

Pouch FillingPouch FillingPouch FillingPouch FillingPouch Filling

Detectors can be installed to inspect powdermaterial prior to filling in preformed aluminum

pouches. Apertures of 50mm (2) or 75mm (3)are usually sufficient but space limitations restrict

the exterior case dimensions. Auger fillers aregenerally used, with the special zero metal freezone (ZMFZ) detector installed between auger and

8/12/2019 Red Guide

20/3316


IIIIIt is normal for some time lag to be required fromthe moment of detection to the moment of rejectionto allow the contamination to move to the reject

point. This can range from fractions of a second onhigh speed applications, where the detector and rejectdevice are close together, to as much as 30 seconds

when rejection, either manually or automatically, isplanned at some remote point. A second independenttimer is also required which will control the length oftime the reject device operates. This is usuallyadjustable from about 0.5 seconds to 10 seconds. Theshortest time would be required on a punch type rejectbut a retracting band system would normally operatefor several seconds to remove larger items from a slowmoving belt. Both timers would be available as astandard item from detector manufacturers.

It is important that the timers are immediately resettableand that the detector is still operative while timing out.The detector must be capable of detecting a secondparticle and also resetting or extending the timer toensure it is rejected. A continuous stream of metalparticles should result in the reject device operatingcontinuously until all the particles are removed.

7.1 Variable Speed/Stop-Start Applications

Accurate rejection and timing become more complexif the transport conveyor has variable speed or can bestopped with product between the detector and rejectsystem. The time taken for the product to move to thereject position is not constant and so a simple time

delay method cannot be used.

The normal solution is to use an electronic shift registerwhich can monitor belt movement and the position ofproduct on the belt. A shift register is a device whichwill give an output signal after it has received apredetermined number of input pulses. It is of noimportance if these pulses are received rapidly or overa long period. The input pulses are produced by a pulsegenerator fitted to the shaft of a roller on the conveyorsystem. It is normally made from a metal disc withteeth or holes cut into it. Figure 31 shows a typicaldevice. Each time a tooth of the disc obscures the

photoelectric device or passes close to the proximity

sensor, a pulse is generated. In this example, eachrevolution of the disc generates thirteen pulses. Ahigher number of teeth will produce more pulses perrevolution, which ultimately will allow a more accurateregistration of the product.

The signals from multiple or consecutive metal particles

will be stored in the shift register and sequencedthrough, ensuring each one is subsequently rejected.

When installing systems in a line which may operate atvarious speeds, it is not always necessary to have thedetector system also variable. The additional expenseand complexity can often be avoided by fixing thedetector system to run at the normal line maximum.

7.2 Photogated Timing

Figure 32 shows a punch reject, with the timingadjusted to accurately reject the metal particle located

in the center of the pack. If a particle occurs at theleading or trailing edge, the reject may operate too

7. REJECT TIMING

8/12/2019 Red Guide

21/33


1

early or late, possibly missing the pack or disturbingthe neighboring packs and causing a line blockage. Ifusing air blast or diverter arm rejection, a possiblesolution would be to adjust the timers to operate earlyand for an extended period. This, however, wouldremove several good packs and most likely spin ordisturb others.

The best solutionand the only one when punch rejectsare usedis to accurately monitor the position of thepack and operate the reject device when it has reachedthe correct position. The technique is known asphotogat ing. This will ensure accurate rejectionindependent of location of metal in the pack.

8/12/2019 Red Guide

22/3318


MMMMMetal detectors may be used at various stages of aproduction process.Bulk InspectionBulk InspectionBulk InspectionBulk InspectionBulk Inspection

Eliminates metal before it can be broken intosmaller pieces.

Protects processing machinery from damage.

Avoids pro duct and pa ck ag ing wa st e bysubsequently rejecting a finished higher valueproduct.

Typical examples include bulk inspection of meat blocksprior to grinding, ingredients for pizza toppings andgrain.

Finished Product InspectionFinished Product InspectionFinished Product InspectionFinished Product InspectionFinished Product Inspection

No danger of subsequent contamination

Ensures compliance to quality standards

A combination of bulk and finished product inspectiongives optimum protection.

Selecting a reliable metal detection system is just thefirst step in achieving the final objective - minimizingor eliminating the incidence of metal contamination.

Those responsible for establishing and monitoring theprogram should ensure proper procedures are clearlyspecified and implemented and that the line operatorsand general work force are aware of them.

8.1 Sensitivity Standard

Establishing a sensitivity standard can be relatively easyfor producers of small dry items such as confectionery,but more difficult when a wide range of producteffect lines are produced. Agreeing a minimumcompany standard for finished product inspection willhelp overcome the possibility of a detector beinginstalled at the wrong place in a production line. Anexample is where inspection of finished cases is beingconsidered instead of inspecting each individual item.The larger detector would be of lower sensitivity andfrequently occurring metallic specks in the cartonmaterial would, undoubtedly, limit the detectorscapabilities.

Equipment should be operated at its maximum reliablelevel, which may be better than the agreed standardson some applications. It is, however, more importantfor equipment to work reliably, long term, and withoutfalse rejects, than to try and achieve a better sensitivityand create false alarms. For dry non-conductiveproducts, the following standards are used by manyfood processors.

Standards for very wide detectors, such as those usedat the outlet of ovens or enrobers, may need to beslightly lower than this.

For conductive products, product testing is needed todetermine detection capabilities. Test results, however,should only be considered as an estimate.

Both the minimum acceptable standard and the

individual line specification should be determined forboth ferrous and non-magnetic stainless steel. The linespecification should be marked clearly on the side ofthe detector and metal samples of the correct diametersshould be available for testing.

Access to the detector controls should be by authorizedpersonnel only.

8.2 Testing the Detector

Testing the metal detector with known test samplesshould be carried out regularly to confirm both

detection and accurate rejection. The interval betweentests will depend on the implications of a failed test(see 8.3). Every 2-4 hours is typical. It is important tofollow the correct procedure.

Use correct test samples in ferrous and stainlesssteel mounted in a plast ic card or plastic block.

For discrete items, place the metal sample at thefront of a well identified product and pass throughthe detector, ensuring it breaks the photogate, ifused.

For bulk product, place the sample in the productflow.

Confirm the product with sample is detected andrejected.

For discrete items repeat the test with the sampleat the back of the product.

A more demanding test is to use two samples, correctlyspaced, to ensure multiple rejection.

8. ESTABLISHING AN EFFECTIVE METAL DETECTION PROGRAM

8/12/2019 Red Guide

23/33


1

Precautions must be taken to ensure that any testsamples not rejected do not become lost on theproduction line. If the detector is positioned just priorto another processing machine, such as a grinder ormixer, it is wise to attach the test sample to a piece ofstring.

8.3 Action Required if Test Fails

If the test sample is not detected or rejected, all outputsince the last successful test should be consideredsuspect and possibly contaminated. The cause of failureshould be determined and if a result of tampering or achange in production conditions, procedures shouldbe established to prevent reoccurrence. If the unit canbe adjusted to bring it back to correct operation, thisshould be done and noted on the test log.

If a test shows that a system is faulty, the user mustdecide what course of action to take. Several optionsare available:

Continue production and repair the system as soonas possible.

Continue production and re-inspect productthrough an off-line system.

Stop production.

Stop production and re-inspect everythingproduced since the last successful test. If this isnot possible, quarantine warehoused product.

Only by stopping production and preferably re-inspecting product, can a company be confident thatthey have an effective metal detection program in place.

A corporate policy should be agreed and publicizedahead of time. The value of taking this position isevident in the event of a complaint or possible litigation.

The temptation to keep production lines running canbe reduced if the metal detector is designed with anelectronic module which can be easily swapped out bythe user without skill or specialized equipment.

8.4 Treatment of Rejected ProductAll rejected product should be reinspected in the samedetector during a break in production or in a separateoff-line detector, to locate the offending metal piece.For discrete items, the following procedure issuggested.

The contaminated items should be passed threetimes through the detectors in various orientations.If there is no detection, the item can be consideredacceptable.

If the item is detected, remove any packing materialand repass.

Divide product into smaller and smaller pieces untilmetal can be located.

Do not try and locate metal visibly by spreadingproduct on a tableyou wont be successful.

Metal particles found should be shown to line personnelso they build up confidence in the equipment and thenkept for future reference and details recorded in aMetal Contamination Daily Log. If the source isknown, it should also be recorded. If not, investigationsare extremely useful in preventing a reoccurrence andcan result in a change in maintenance procedures oreven a change in raw material suppliers. Locating andretaining the particles has the added advantage that ifa screen or blade, for example, is known to have brokeninto the product, the individual pieces detected can becollected and the component reassembled to ensure

nothing has been missed.

8.5 Performance Validation

A metal detector with a Performance Validation Routine(PVR), can help ensure testing is carried out properly atagreed time intervals with proper test samples and, ifrequired, provide hard copy documentation confirmingthe test.

PVR will call for testing at an agreed time interval.Approved personnel enter a personal access numberinto the detector to allow the test to be completed

with the correct sample size. Failure to test theequipment at the agreed time interval can action avariety of outputs.

Hard copy documentation showing that testing hasbeen carried out can be provided through a local printeror downloaded to a central PC using a detector networkcapability.

Proof of Inspection is useful evidence in the event ofa customer complaint and to confirm compliance toan inspection standard in a Vendor CertificationProgram.

8.6 Detector Networking

Microprocessor based metal detectors may benetworked together and linked back to a central PC.This provides the user with two important benefits:

Real Time StatusReal Time StatusReal Time StatusReal Time StatusReal Time Status

By glancing at the PC screen, a supervisor may beimmediately reassured that all detectors areoperating properly and have been recently tested.Warning flags can be displayed and corrective

8/12/2019 Red Guide

24/3320


action immediately taken if a sudden increase inrejections occur, if a detector fails, or if metal isdetected but the reject device fails to remove it,for example.

Accurate Documentat ionAccurate DocumentationAccurate Documentat ionAccurate DocumentationAccurate Documentat ion

All events such as rejections, faults, setup changesand QA tests, are stored in a database with the

time and date they occurred. The information maybe viewed or printed in various report formats. Inthe event of a complaint, a printout can prove the

product was inspected by a detector which wasproper ly adjusted and confirms when and whotested the unit with test samples. Additional shiftreports can prove compliance to inspection

standards and help monitor HACCP improvements.

It is quite common, when detectors are initiallynetworked, for several flaws in the overall metaldetection program to be highlighted within the first

24 hours. These can then be corrected, thus improvingthe overall protection level.

8/12/2019 Red Guide

25/33


2

FFFFFrom the authors experience, relatively fewincidents of metal being undetected are a resultof the detector failing. They are usually associated

with poor working methods by company employees andincorrect system design. Often the complaints do notresult from tiny metal pieces but from larger items such

as washers, bolts, and pieces of blades and screenswhich should be detectable by even the most basic typeof detector.

There will always be a finite limit to the smallest metalpiece detectable and processors should ensure that thislimit is understood and acceptable to them.

9.1 Conveyor Design

The design and method of manufacture of the conveyorwill have a great influence on the detector. It is anintegral part of the complete inspection system and, in

almost all cases, the detectors maximum sensitivitycannot be achieved when fitted to an existing generalpurpose transport conveyor. Fully welded structures,incorporating correct metal free zones and properlyisolated rollers, pulley, cross structures and detectorhead mounting are essential to obtain the highestreliable performance. Conveyor belting must be metalfree to a very high standard and suitably jointed.Antistatic belting should be avoided.

If these problems are not solved at source, the commonoutcome is a gradual increase in false rejects. The easysolution is to reduce the sensitivity control of thedetector, sometimes to an unacceptably low level.

9.2 Non-Positive Reject System

This is probably the weakest link in of the wholeinspection system. There are a number of commonproblems. These include:

Reject not suitable for the application.

Reject not photogated on long discrete packs.

System design not capable of removing consecutivecontaminated packs.

Failure of the reject due to low air pressure,blockage or solenoid failure.

Downstream product backup through the detector.

Product spacing and reject design not compatible.

One the benefits of single source responsibility forconveyor, reject and metal detector, is that these issuescan be addressed at the design stage, if necessary.

The author recalls one instance with a frozen pizzamanufacturer using eight (8) identical detector systemswith air blast rejection but no potogating. Theequipment was tested each hour by placing the testsample in the center of the pizza and each time thetest was successful. It was then suggested to repeat

the test with the test sample on the front edge of thepizza. The QA manager was amazed to see the pizzain front of the sample rejected and the contaminatedsample continue down the production line. The reasonfor his ongoing customer complaints was clear.

Simple additional control devices are available to ensurea reject device is operating properly and thatcontaminated packs are accurately rejected.

9.3 Production Continuing When theSystem is Faulty

A firm policy should be established dictating the actionrequired in the event of a detector fault. Haltingproduction is the safe option. Alternatively, productcan be stored for subsequent inspection in an off-lineunit prior to release. If your detectors have a userreplaceable quick change module, a repair can beeffected in minutes and the temptation to keep runningavoided.

The drive motor of the conveyor can be wired to stopin the event of detector failure or the reject device madeto operate continuously. Automatic self checking is anessential feature of any metal detector. Its purpose is

to continually monitor the detector and give an alarmif it fails or if the sensitivity falls. This automaticchecking can be extended to cover all associated itemssuch as reject mechanisms.

9.4 Reject Product Returned to Production

If product is rejected onto the floor or into an opencontainer, it can be easily returned to production inerror or, when production schedules are critical,intentionally. Rejection into lockable reject bins helpsovercome the problem. A warning device should be

incorporated to indicate when the bin fills up. Frequentfalse rejects and erratic operation can undermine theoperators confidence in equipment so it is importantto ensure these are avoided. In addition, showing lineoperators the various metal pieces found will buildconfidence in the equipment.

Most microprocessor equipment will display thenumber of rejected items on the control panel and thisshould be verified with the actual number found. Goodrecord keeping will highlight which lines or shifts seemto have suspiciously few rejects.

9. REASONS WHY YOUR PROGRAM MAY FAIL

8/12/2019 Red Guide

26/3322


9.5 Line Operators Working Practices

To prevent the detector being switched off or reducedin sensitivity, access to the controls should be restrictedto Q.A. personnel or to a responsible supervisor trainedin the adjustment of the machine. This is best achievedby a password access code on digital systems.

Other problems witnessed by the author include air

supply to reject device disconnected, reject bin overfulland reject arms tied open with string. Additional simplechecking devices are available to monitor these.

9.6 Conveyor System Used as a Pack OffSystem

If all product is to be removed manually from theconveyor, the reject device should be positioned as closeas possible to the detector and the space betweencovered with a clear guard to prevent contaminateditems being removed manually, in error, before theyhave arrived at the reject point.

9.7 Subsequent Contamination AfterInspection

For quality assurance applications, the ideal point ofinspection is immediately after packing or as close tofinal packing as possible. When the packaging materialincludes aluminum foil, there is an option to check witha normal type detector before packing or after packingwith a ferrous-in-foil type detector. However, asstainless steel and non-ferrous metals are not detectablein the ferrous-in-foil unit, this is only recommended

when no alternatives are available.

Intentional inclusion of metal is particularly difficult toprevent. Inspection as late as possible in the processand minimizing the access to finished product will help.Employees under notice of termination should not beallowed in sensitive production areas. If sabotage is ofparticular concern, discussions with the metal detectormanufacturer are likely to prove worthwhile.

9.8 Narrow Zone or Zero CrossoverTechnique

Multiple metal particles frequently occur in practice.Detectors using the Narrow Zone technique aresusceptible to being fooled under these conditionsallowing metal particles to pass undetected. A fulldescription of these problems are given in Chapter 3.

9.9 Users Unaware of Detector Limitations

Most users are unaware of the practical limitations ofthe detectors they use. A detector working at a 2mmsensitivity, for example, will not necessarily detect allmetal pieces larger than 2mm. A thin piece of screenwire could be 25mm (1) long or more and still beundetected. An understanding of the orientation effect(see section 4.3) will prevent a false sense of securitydeveloping and this can influence the location of thedetector in the production process preventing it frombeing installed to perform a task for which it is

incapable. Even slight reductions in operating sensitivitycan have a significant effect on the performance of theequipment, a point seldom appreciated by the user. Ifa specific type of contamination is common, this shouldbe discussed with the detector manufacturer during asite visit, together with all other relevant information.

9.10 Detector Drift

Detector drift occurs over a period of time as a resultof temperature and humidity variations, aging ofelectronic components and buildup of product in the

aperture, resulting in changing sensitivities and falsealarm or nuisance signals. Typically, the sensitivity toferrous metal improves with drift while non-ferrous andstainless steel becomes worse. For this reason, it isalways important to test the detector with both typesof metal or the drift problem may be masked. QuartzFrequency Control and Automatic Balance Control willgo a long way towards eliminating drift, ensuringsensitivities are maintained permanently.

8/12/2019 Red Guide

27/33


2

MMMMMoney spent reducing complaints inevitablyyields a better return than money spentanswering them. The real value of a quality

program is determined by its ability to contribute toprofits and to customer satisfaction. Any expenditurewhich could have been avoided will have a direct, but

negative, effect on profits. The value of clearlyidentifying the cost of quality should be obvious.

An argument for quality improvements is often weakwhen it has to deal in generalities and opinions. Itbecomes more convincing and realistic if it can quantify,in money terms, the costs and savings. These costs cangenerally be split into three broad areas.

Prevention CostsPrevention CostsPrevention CostsPrevention CostsPrevention Costs

These cover the activities specifically designed toprevent contamination or defects. They shouldcover items such as supplier capability surveys,

employee education and training and establishinggood manufacturing practices. They are extremelycost effective.

Appraisal CostsAppraisal CostsAppraisal CostsAppraisal CostsAppraisal Costs

These cover costs associated with test ing,inspection and ongoing evaluation of the

production process, to ensure conformance withquality standards.

Failure CostsFailure CostsFailure CostsFailure CostsFailure Costs

These costs, potentially by far the highest, cover

failure occurring both before and after shipmentof product. A metal contaminated product foundbefore shipment is a failure resulting in productand packing wastage, possible machinery damageand loss of output. If discovered after shipment,loss of customer satisfaction, product recall, adverse

publ icity and potential law suits can result.

Prevention and appraisal costs are generally low andare incurred because poor quality may exist. Failurecosts are high and are incurred because poor qualitydoes exist.

A systematic approach needs to be developed andintegrated into the overall Q.A. program. It needs tobe rigorously set up and rigorously followed. It needsto be pro-active rather than re-active and used toultimately prevent the occurrence of contaminationrather than just detect it. Quality responsibility shouldbe passed back to include suppliers, to ensure theirstandards are equally demanding. Often contaminationin the suppliers product is more easily detected beforeit is further processed and broken into smaller pieces.

The objective should be to have control over the wholeproduction process, the incoming raw materials, theenvironment, the processing and the packaging. As a

first step, a Foreign Material Task Force should beformed to develop, implement and coordinate theforeign material control system. Ideally, it shouldinclude senior personnel from production, qualityassurance, engineering and maintenance.

The test force has three main responsibilities:

Establishing and monitoring Critical Control Points

Developing Good Manufacturing Practices

Providing Documentation and Trend Analysis.

10.1 Critical Control Points

Hazard analysis and establishing critical control points(HACCP) is an important first step in taking a pro-activestep towards reducing contamination. A flow diagram

should be drawn up showing the traffic of all productsin each production stage, from the incoming rawmaterials to the final warehousing of finished product.Depending on the complexity and number of differentprocesses, up to ten separate flow diagrams may berequired. Every point in the product flow should beconsidered and those which may create a potentialhazard are identified as a Critical Control Point (CCP).The CCP is coded and the method and frequency ofchecking is determined. Control points can beestablished not only where metal contaminationhazards are possible, but also any other quality relatedmatter.

To illustrate how this can be developed in practice, anexample is shown for the production of breaded fish

portions.

Example:Example:Example:Example:Exa mpl e: The complete manufacturing processinvolves a number of distinct stages:

Fish coating manufacture and storage Frozen ingredient supply and storage Processing breaded f ish pieces Packaging breaded fish piece

Figure 35 shows the flow line for the processing stage.Each point on the flow line where a potential hazardfrom metal contamination exists, is identified andcoded. A description of the potential hazard is drawnup with details of how it should be checked and bywhom. Figure 36 shows the hazard control list. Thesehave identified hazards from metal contamination butthe system can be expanded to cover all potentialhazards, such as microbiological contamination, deliverytruck temperature, cooking dwell time and portionweight control.

A separate reporting form should be drawn up for eachcontrol point to confirm that the necessary procedureshad been carried out.

10. DEVELOPING A FOREIGN MATERIAL CONTROL SYSTEM

8/12/2019 Red Guide

28/3324


8/12/2019 Red Guide

29/33


2

10.2 Good Manufacturing Practice

In addition to identifying Critical Control Points, otherpotential sources of contamination, not related directlyto the production process, need to be identified andprocedures introduced to eliminate the danger.

These could include:

Specifications for raw materials should state thatthey are free from foreign body contamination and

should indicate specific precautions for the supplierto take, depending on product type, such as

powder material to be screened and passed througha detector, carcass meat not to be labeled withmetal tags, no stapled container to be used.

Paper clips should not be used on documents inproduction areas.

No thumb tacks should be used on any noticeboard.

No hair c l ips, watches, jewelry a l lowed inproduction areas.

Protective clothing should have no outside pockets.Laundered items to be checked for loose buttons

prior to re-issue.

Only magnetic Band-Aid wound dressings to beused by personnel, to aid detection of lostdressings.

Conveyor lines carrying open containers should be

covered until the containers are closed or capped.

Holding containers should be covered.

Step bridges over production lines should haveenclosed sides and be checked regularly.

Many other specific effective measures can be carriedout relevant to specific industries and manufacturingprocesses.

Effective training is fundamental for all involved, fromthose who design plant layout to the unskilled

production line operator, to have an awareness andcommitment to the avoidance of foreign bodies.

Programs should include an explanation of thecompanys Quality Assurance philosophy and details ofthe Critical Control Points. Documentation andreporting procedures should be detailed and individualresponsibilities for reporting potential hazards such asdefective machinery, made clear.

Poor working practices during maintenance is acommon cause of metallic inclusion. A routine andpreventative maintenance list, which can be conductedoutside normal production hours, should be drawn up

and regularly updated. Other effective proceduresinclude:

Equipment maintenance should not be carried outduring plant operation, particularly welding anddrilling. Screens should be provided to prevent

spread of welding slag and swarf.

For major work or new installations, complete floor

to roof screens should be used.

Magnetic mats, brushes and vacuum cleaners,shou ld be used fo r cleanup and on repa ir edequipment, prior to returning to the productionarea.

Star tup team re sponsibi li tie s, wi th pa rti cu la remphasis on Critical Control Points, should bedetailed.

On completion of any repairs or installations, theplant and surrounding area should be inspected bya member of the Q.A. team.

Implementing proper procedures and working practicescan help the quality philosophy permeate the wholecompany. Appropriate channels for rapid feedback andrevision, in light of new experience, will keep thesystems live.

Every piece of metal which is prevented from enteringthe production process represents a 100% success. Nodetection system can ever reach this level.

10.3 Documentation and Trend Analysis

The effectiveness or otherwise of monitoring CriticalControl Points, can only be determined by efficientcollection of data and trend analysis.

The importance of recording each incidence of metalcontamination has been described in Chapter 8. Trendanalysis of contamination type and frequency, line byline, or machine by machine, can identify particularsources of trouble such as raw material supplier,production staff or shift, or inadequate maintenance.Using this information over a period of time will helpdetermine the effectiveness of the Q.A. program and

equal ly important, wi l l be the f i rst step inquantifying--in monitory terms, the savings or increasedprofit generated. Monitoring the critical control pointsshould result in a significant reduction in quite a shortperiod. Two useful trend charts record the number ofpieces of metal contamination detected on a weeklybasis and also the number of consumer complaints.Each incident should be investigated to determine ifthe failure was a result of ineffective monitoring of theCCP, if a new previously unidentified CCP is responsibleor if the metal particle is smaller than the operatingcapability of the metal detector.

8/12/2019 Red Guide

30/3326


DDDDDespite the widespread use of metal detectorsover recent years, there are few guidelinesavailable to help users evaluate a detector or

compare the capabi l i ty of different brands ofmachines.

Those new to metal detectors may draw up a check

list of features between different brands and makean assumption that the brand with the longest list isthe best choice. Frequently, they may also ask Whichunit is more sensitive, as a basis to evaluate andcompare. Neither method is likely to yield helpfulguidance.

One characteristic of microprocessor technology is thatunlimited features can be provided but these featureswill not necessarily contribute to a metal detectorsoverall effectiveness. Users with more experience willalso know that sensitivity is only one factorand almostcertainly not the most important.

Drift, erratic detection, complexity of setup and randomfalse rejects are the key factors which will make adifference in the success or failure of the overall metaldetection program. Metal detectors can be frustratingto production personnel when they seem to operateinconsistently. They will quickly loose confidence in aunit that rejects product which is subsequently shownto be good or one that requires constant attention forthe sensitivity standard to be maintained.

What then are the practical factors which potentialusers should consider when selecting a detector brand?

The following are those considered most important byusers with long term metal detection experience.

11.1 Stability

This is thedistinguishing factor of a top quality metaldetector and highlights the difference betweensensitivity and performance. Performance is ameasure of equipment capability under real plantconditions.

A stable metal detector is able to operate consistentlywithout false rejects or erratic detections and should

not require periodic adjustment. Most microprocessorunits will give similar sensitivity levels when tested sideby side under laboratory conditions. However, overextended operation on a production line, significantdifferences will become evident.

As sensitivity levels increase, a detector will graduallybecome more unstable until an unacceptable level isreached. Electronic drift is one cause of instability andoccurs over a period of hours, days or weeks. AutomaticBalance Control is one technique used to minimize drift.

The detector capable of giving consistent reliabledetection, without the frustration of false rejection, will

win the confidence of both l ine operators andmanagement and provide the best protection longterm. Production line performance is the measurewhich takes all these factors into account.

An unstable detector, particularly when linked to anautomatic reject device, can quickly become a focus of

criticism.

11.2 Repeatability

In addition to false rejections, drift can cause thedetection level to vary over time. The effect ishighlighted with stainless steel samples and for thisreason, stainless should be used for testing. Having adetector that detects the test sample repeatedly eachtime it is passed, over a period of weeks or months,instills confidence in the user. It also avoids theproblems of having to rescreen product or allowing

metal to pass undetected.

11.3 Ease of Setup

A detector which has a complex or confusing setupprocedure inevitably will not be adjusted correctly. Evenso called Auto Setup units will require programming.After initial instruction, it should be practical for theuser to adjust all parameters without reference to aninstruction manual. A logical procedure avoids havingto memorize special sequences and will mean changescan be properly made long after the initial instructionis given.

11.4 Washdown Integrity

For producers of meat, poultry, dairy and similarproducts, a detectors inability to withstand frequentheavy duty washdown is a common problem. Repairof a leaking detector is both expensive and timeconsuming. If a detector needs to be removed fromline for repairs, your program may be compromised.Typically, there are two weak points:

The joint between the stainless steel case and theplastic l iner through the detector opening.

Thermal expansion caused by temperaturefluctuations during use and washdown, coupledwith the difficulty of bonding to stainless steel,allows this joint to crack, permitting moisture toenter. Premature failure is often the result. Somedetectors have eliminated this weak point bycovering the plastic/stainless joint with an additionalbolted flange, which will withstand the harshesttreatment for extended periods.

11. COMPARING METAL DETECTOR BRANDS

8/12/2019 Red Guide

31/33


2

Leakage around the control cover.

Detectors equipped with a sealed washdown coverover the controls, at some point will be left openor not tightened during cleanup. Frequent openingwill also eventually wear the gasket seal. Somedetectors avoid this problem by using a washdownmembrane keypad for setup, designed for directwashdown that requires no cover.

Other factors should also be considered depending onindividual requirements. The importance of conveyor/reject design was discussed in Chapter 9. Detectorsuppliers with in-house system design will be able togive single source responsibility for a successful startup.

On lines where downtime is unacceptable, detectorswith a single electronic module designed for userswapout will be important.

8/12/2019 Red Guide

32/3328


12. FINAL QUESTIONS

IIIIIn the event of a consumer complaint or possiblelitigation, how certain can you be that there are noflaws in your metal detection programand can you

prove it?

A companys metal detection policy should be definedand publicized. Ideally, to minimize exposure to risk, it

should be in line with Figure 37.

In addition, whenever possible, you should ensure yourraw material suppliers adopt a similar policy to ensurethey do not ship you contaminated product. Manymajor corporations now make satisfactory metaldetection a part of their Vendor Certification Programs.

Having such a firm policy is essential to close the loop

but it can cause a serious dilemma at the operatinglevel. In the event of a failure, if production is stoppeduntil a factory trained technician can schedule a servicevisit, the loss of output will be substantial. Alternatively,if production continues, your metal detection programwill be flawed and you cannot be certain of producinga metal-free product.

One of the great benefits of using metal detectors whichincorporate a universal quick change electronic module,designed for user swapout, is that it not only minimizesservice costs, which may be relatively minor whencompared to the cost of lost production, but, more

importantly, it avoids the temptation to run aproduction line unprotected.

12.1 Equipment Audit

Outside audits of equipment carried out by certifiedquality auditors is an additional service which isavailable to assure users that equipment is incompliance. Experienced metal detection experts canoften spot potential problem areas and suggestsolutions before they become apparent to the user. Thisservice is currently available from Safeline.

ABC COMPANY

Metal Detection Policy

1. All our products will pass through a highperformance metal detection system.

2. All detectors are checked by qualifiedpersonnel every ___ hours, with agreed metal

test standards. Documentation is producedautomatically indicating when the test iscompleted, with the result.

3. If the detector is not tested within the agreedtime period, an alarm warning is given andthe production line is stopped automatically.

4. If the detector fails a test, the production lineis stopped and all output since the lastsuccessful test is quarantined.

Fig 37

8/12/2019 Red Guide

33/33

SAFELINE LTD

510 Montford Court

Montford Street, Salford, M5 2SN, EnglandTel: 161 848 8636 Fax: 161 848 8595

SAFELINE INC

Safeline Business Center

6005 Benjamin Road

Tampa, Florida 33634 USA

Tel: 813 889 9500 / 800 447 4439 Fax: 813 881 0840

Service: 800 259 9093

World Leader in Metal Detection

METAL DETECTION


Red Guide

Documents