GB2131161A - Measuring reflectance, handling, and sorting articles such as furs - Google Patents

Abstract

Apparatus for optical measurement of a surface, especially for measuring the hue and lightness of fur, comprises means for applying suction to a portion of the surface, eg to clean and raise the fur, and means for measuring the reflectance of a portion of the surface that has just been subjected to the suction, eg a light source and a photosensitive cell (with a filter wheel if necessary) aligned at a selected angle to the direction of the fur. Fur pelts are transported on a conveyor to a measuring station where a reciprocating carriage moves a suction nozzle having a knife edge to raise the nap across the pelt. The source, a rotating filter wheel with blue, yellow and white filters, and the detector, which is angled to avoid specularly reflected light, are <IMAGE>

Description

SPECIFICATION Improvements in and relating to the handling, examining, and sorting of articles The invention relates to the handling, examining, and sorting of articles, and especially to the automatic measurement of reflectance of surfaces, especially of fur, for example, mink fur, the preparation of fur for such measurement, and the sorting of pelts according to the results of such measurements.

The invention provides apparatus, devices, and methods as claimed in the appended claims.

In accordance with the present invention, in its various aspects, it is possible to construct an apparatus to which animal pelts roughly sorted by size and colour type may be fed and which will automatically prepare, assess, and sort these pelts according to their exact shades.

One form of machine for assessing the shade of mink pelts, constructed in accordance with the invention, will now be described by way of example with reference to the accompanying drawings.

Overall Description of Automatic Colour Evaluator Introduction The Automatic Colour Evaluator for mink pelts consists ofthree elements; the inspection module (Figure ), the distribution module (Figure ) and the take-off conveyors (up to three in number). Within the inspection module the pelt is first prepared and then viewed to evaluate its shade. The distribution module segragates the pelts according to the results of the evaluation and feeds it into the appropriate take-off conveyor.

The evaluator has been designed to process pelts at a rate of up to 40 per minute and requires a maximum of five operators; two for loading the machine and one for each take-off conveyor. The maximum size of pelt acceptable to the machine is 1.1 m long (including its tail), 0.1 sum wide and 30mm thick.

The inspection module and the distribution module combine to give an overall machine length of 4.2m with each module being approximately 1.7m high and 1.4m wide.

Pelt Transport The pelts are incremented through the machine by two walking beam mechanisms giving a step length of 170mm. This represents the distance between successive stations in the inspection module where operations may be performed. It is also the width of each of the nine trap doors of the distribution module (through which the sorted pelts leave the machine). These mechanisms have a series of fingers which push the pelts along the table of the machine from one station to the next. One is situated beneath the table of the insepction module and the other is above the table of the distribution module.

Machine Control Two computers are employed in the machine control system. The master computer supervises the sequencing of the machine whilst the slave computer is responsible for the evaluation of the pelts' colour.

The Colour Evaluation Process Initially the machine is set for the size of pelts to be processed. The pelts are then unloaded onto a table at one end of the inspection module. The operator positions the pelts the right way up, with their noses against a guide rail, and their sides against a second guide which ensures that the pelt is in front of the first set of feed fingers on the walking beam. This procedure for handing pelts is similar to that already employed with the mink heads towards the operator and their direction of transfer through the machine from left to right.

From the load station the walking beam pushes the pelt under a height gauge. This gauge, set at 30mm, is necessary to protect oversize pelts from damage by the moving elements of the machine. Pelts that are prevented from entering the machine by the height gauge cause the feed fingers to deflect downwards and so leave the pelt on the loading table for the operators to remove.

The next two stations of the machine are not used; they are reserved for additional pelt preparation should this prove necessary.

The third station on the machine is the brushing station. Here the pelt is clamped by vacuum to the table top and three brushes, mounted on a belt, brush against the nap to remove debris from the pelt and any irregularities in the fur. Here the measurements are taken which enable the shade of the pelt to be assessed.

A pad set into the table top lifts the pelt to press it against a mask. The strip of fur visible through the aperture of the mask is first vacuumed to remove dirt, then illuminated with light from a stabilised source. The reflected light is measured by a photodiode on a reciprocating carriage as it passes over the strip of pelt. The output from the photodiode is analysed by the measurement computer and associated electronic systems to decide the shade of the pelt. The master computer takes this information and allocates the appropriate door in the distribition module to the pelt.

When the measurements have been taken the pelt is lowered back to the table in time to be moved on by the walking beam. The next two movements of the walking beam bring the pelt onto the transfer station which links the inspection and distribution modules.

Pelt Distribution Cycle Beneath the distribution module are three take-off conveyors each one segregated into three independent tracks.

The pelt is indexed down the length of the distribution module from the transfer station by the walking beam mechanism situated above the table top until it rests on the correct door for ejection. The ejection process, which is initiated by the master computer, consists of the door beneath the pelt swinging down so that the fingers on the walking beam can push the pelt forward and out onto the conveyor track. The door is then closed before the next movement of the walking beam. The pelt which has been ejected is carried away by the conveyor to an operator who places it in the appropriate trolley for its shade.

Mechanical Details Introduction A general description of the automatic colour evaluator is given in chapter five. In this chapter the mechanical aspects of the machine are discussed in greater detail.

Machine Framework The automatic colour evaluator is based on two main frames, these being: (a) The Inspection frame (b) The Distribution frame Inspection Frame The frame, shown in Figure 6.1, is made of two sub-assemblies: the lower frame and the upper frame.

Both structures are made of rectangular hollow sections (RHS) 60 x 40 and 40 x 40 mm, and supported by four swivelling caster wheels. The closed box arrangement of the lower frame houses the cradle and the two electronics cabinets. The top of the lower frame supports the table top and the upper frame.

Cradle (Figure 6.2) This unit is a sub-frame which provides the supportforthe motor and the walking beam mechanism. The cradle slides into the lower inspection frame on three flat steel strips. When the cradle is positioned, its height is adjusted by making suitable spacers. The frame is then drilled and tapped so that the cradle may be bolted into its final position.

The Table Top This assembly of flat plates and angles is made of anodised aluminium. A pair of longitudinal slots is provided for the 'fingers' of the walking beam. These slots are normally closed by brush strips which stop hair and dirt from falling into the mechanism located below the table.

The Upper Frame The purpose of this sub-assembly is to provide a structure which supports the brushing and viewing stations, and provides anchoring points for the drive system including the shaft which transmits the drive to the distribution module.

Distribution Frame (Figure 6.3) The distribution frame is designed to span the three conveyor belts which carry the pelts away after sorting. The structure is made of 60 x 40 and 40 x 40 mm rectangular hollow sections and 40 x 40 mm angles. This frame is supported by four swivelling wheels for ease of maneouvre when the distribution and inspection frames are being linked together. The linking is achieved with a pair of couplings which allow adjustments in translation and rotation during the operation.

The distribution frame carries the door control mechanism and the doors. On the top of the distribution frame, provision is made to accommodate the second walking beam mechanism, and the door drive control shaft.

The drive unit consists of a three phase induction motor with integral brake, connected to a gearbox through a variable speed belt drive. It can be adjusted manually to give an output speed range of 20 to 60 rpm.

The drive arrangement is illustrated in Figure 6.4.

The output from the integral reduction gearbox first passes through a torque limiter which can be set to disengage the drive at chosen output torque to protect both the operator and the mechanisms within the machine. From the torque limiter the drive is split in two directions by a T - gearbox.

The lower output drives the walking-beam mechanism of the inspection module and incorporates an additional shaft for the connection of a handle so that the machine can be 'turned over' by hand for setting purposes. For ease of maintenance and setting, the motor, torque limiter, gearboxes and walking beam mechanism are mounted on a removeable cradle which is located in the inspection module. The cradle assembly is shown in Figure 6.2.

The upper output is extended by a shaft, vertically, to another T - gearbox. One output from this gearbox drives the digitiser via a timing belt. The digitiser is used by the control system to detect the rotational osition of the drive shaft. The other output is used to drive the brushing station and the Cardan mechanism sia a horizontal propshaft and toothed timing belts. The 1:3 speed ratio necessary to drive the pulley on the brushing station is achieved with two timing belts and the associated intermediate double pulley is located Dn an extension of the main Cardan shaft.

The drive for the distribution module is a continuation of the propshaft. A polarised flexible coupling is used at the junction of the two modules to ensure the correct phasing of all the elements of the machine, on assembly. The propshaft within the distribution module drives the walking beam mechanism via a rightangle gearbox. Another toothed belt is used to drive the door closing mechanism from an extension of the main drive shaft of the walking beam, Figure . Throughout the drive arrangement torsionally soft flexible couplings are employed to accommodate any misalignment between the drive shafts. All the gearboxes contain helical bored gears for efficient and quiet power transmission.The motion required to index the pelts through the machine is provided by a rectangular motion Walking Beam mechanism and this system has the following advantages: (a) Different cams could be fitted to change either the timing, the stroke, the acceleration or the velocity of any combination of these features.

(b) The mechanism was simple in operation.

(c) It would be easier to engineer due to the relatively small number of parts.

(d) The principles of the system had been well proven in many other applications.

(e) Maintenance would be minimal.

Design of the Walking Beam Mechanism The two modules of the machine required separate Walking Beam mechanisms but they had to be synchronised in order for the pelts to be transferred between the modules successfully.

The square motion which was required is illustrated in Figure 6.5. This was achieved by mounting two parallel beams on a carriage which moved horizontally. This horizontal motion carriage was incorporated inside a vertically moving main carriage. Each carriage was actuated by a separate cam and the camshafts were synchronised by the use of a toothed timing belt. The camshafts were mounted between two heavy steel sideplates which were fastened rigidly into the machine frame. The Inspection module walking beam mechanism may be seen in Figure 6.2.

The principle of operation of the Walking Beam is shown in Figure 6.5. Figure 6.5a shows the fingers about to move horizontally to the right. This action indexes each pelt by one pitch through the machine. In Figure 6.5c the fingers have moved vertically down to clear the underside of the pelt. The fingers then move horizontally one pitch to the left followed by a vertical movement upwards and hence a return to the initial position.

The maximum pelt width that the machine can handle is 150 mm so the discrete pitches created by the Walking Beam mechanism are 170 mm. This dictates the width of the work stations. All subsequent aspects of the machine design e.g. Viewing Station, Brushing Station and Trapdoors, are based on a 170 mm module.

The Overhead Walking Beam Mechanism The second Walking Beam mechanism is used in the Distribution Module and is virtually the same as its counterpart in the Inspection Module. Similar parts are used in the two mechanisms mainly to reduce the manufacturing costs but also to facilitate assembly. The main difference between the two mechanisms is that the Distribution Walking Beam has to be inverted and suspended overhead to allow the unhindered passage of the pelts from the machine.

The Beam Assemblies The Beams in both mechanisms are arranged as ladder-type assemblies (Figure 6.6). The Walking Beam fingers in the Inspection Module are designed to slightly lift the leading edge of the pelt in order to prevent unwanted motion during its transfer in the machine (Figure 6.

There are two potential problem areas: (a) If there was a gap between the fingers and the edge of pelt at the beginning of the cycle, the fingers would subsequently 'hit' the pelt and cause it to move in an uncontrolled fashion.

(b) The pelt moved across the axis of the machine during transfer of the centre line if the pelt was not at 90 to the axis of the machine.

These problems were overcome by the design of curved fingers fitted to the Inspection Module Beam and by incorporating an offset in the Beam Assemblies to ensure that the fingers pushed the pelt with its axis perpendicular to the axis of the machine. See Figure 6.8.

The accurate positioning of the pelt at each work station in the Inspection Module is important so that the colour evaluation is carried out on the same area of each pelt. If a pelt is processed twice it must be evaluated at the same colour both times. Because of the slight variations of colour inherent in any pelt, this can only be guaranteed if the same area of the pelt is evaluated each time.

The positioning of pelts in the Distribution Module is less important because providing the pelt is positioned over the appropriate trapdoor it can be ejected successfully from the machine. For this reason, straight fingers are used in the Distribution Module.

Adjustment of the Inspection Module Beam Assembly The pelts are not of uniform size or shape. However, they will be supplied to the machine after being sorted into large batches of similar sized pelts. An adjustment has been designed into the Inspection Module Beam Assembly so that the discrete pitches of the Walking Beam can be altered with respect to the work stations. This adjustment will be made between batches, if the batches consist of different sized pelts. In this way the centre line of each pelt in the machine will be correctly positioned at each work station.

The Viewing Station Pad The pelts have to be clamped in the viewing station with an area of fur exposed for evaluation. The viewing head which is described in section 6.6.2, cannot be moved down to the pelt so the pelt has to be lifted up into the viewing station mask. The mask which is described in section 6.6.4. also acts as the top half of the clamp.

The bottom half of the clamp is a rubber pad on which the pelt is lifted vertically up into the Mask where it is sandwiched between the two. The rubber pad absorbs any variation in pelt thickness and projecting features such as feet are pressed into the rubber by the mask.

The vertical motion of the pad is actuated by a cam and lever mechanism and incorporates precision shafts and ball bushes as guides. The cam timing is arranged to allow maximum time for colour evaluation and has to be accurately synchronised with the movements of the walking beam. This eliminates the need for a timing belt and reduces the number of shafts, bearings etc.

The pad mechanism is cantilevered off one of the sideplates of the inspection module walking beam mechanism.

Brushing Station Design Introduction Brushing the mink pelt is a necessary step in the preparation of the pelt before any optical analysis. This is because the nap of the pelt must be raised in order for the under fur, as well as the guard hairs of the pelt, to be viewed. Also, a substantial quantity of dirt (sawdust and wool) is present in the pelts and that this can affect the light readings. For successful and consistent colour analysis results, the process of brushing the pelts before optical analysis is necessary. A linear belt brush is used, in which a length of conveyor belt is partly covered with bristles, whilst the rest is left plain. This scheme also has the advantages that: 1. The brushing station can be timed easily to match the machine's sequential operating cycle.

2. The brushing station does not have to be raised and lowered to allow the pelt to be transported into and out ofthe loading station.

The synchronisation of the brushing station with the rest of the machine cycle is critical and therefore a positive method of driving the brushing station is needed. To achieve this, a slatted chain conveyor is used, with the chain sprockets being positively driven from the main machine drive via timing belts. Samples of conveyor chain and wooden brush strips designed to be mounted onto the conveyor chain are shown in Figure 6.11.

Two important criteria in the brushing process are: i) How much dirt would the brushes pick up during brushing? ii) Would the speed of brushing damage the pelts? It was observed that, although the pelt was brushed effectively and was not damaged in any way, the brush bristle did pick up substantial quantities of dirt. This dirt has to be removed from the brushes during the process, to avoid any large clumps of dirt or wool being accidentally deposited on the pelt, which would cause errors to be made in the optical analysis.

Detail Design The brushing station is shown in Figure 6.12. The slatted chain conveyor seen in Figure 6.11 is held between two 16smooth sprockets and three wooden brushstrips are mounted on three consecutive links of the chain. Having brushes on only a short section of the chain allows itto run continuously. The drive to the chain sprockets is via a timing pulley, which is mechanically synchronised with the main drive shaft of the machine. Using the fastest cycle time of the machine - 1.5 seconds - the sequencing of the brushing cycle is shown diagrammatically in Figure 6.13. Allowing 0.5 seconds for pelt transfer, one second is left for clamping, brushing and unclamping the pelt.

The number of links in the chain is determined by the smallest available sprocket size and the desired length of brushing - approximately 300 mm. This resulted in a chain length of 24 links and a sprocket size of 16 teeth (8 effective). Knowing that one rev of the sprocket indexes the chain 8 pitches, 3 revs of the sprocket indexes the chain 24 pitches, or one cycle. Hence a 3:1 drive ratio is required between the sprockets and the main machine drive. For this a drive arrangement using timing belts and pulleys is provided.

As may be seen in Figure 6.11, the two chain sprockets are supported by bearing housings which are mounted between two stiff sideplates. The drive sprocket assembly is fixed, whilst the idler sprocket assembly can be moved in a slot in the sideplates, in order to tension the chain and to facilitate the removal of the chain. In addition, there are two chain support plates which help to prevent the chain being pushed upwards, off the pelt, during brushing. They can also be adjusted to tension the chain.

The problem of "in process" cleaning of the brushes is solved by passing the brushes through a series of combs in the comb box (Figure 6.14), at the end of every cycle. Any dirt or wool present in the brushes is subsequently removed by the combs and drawn away from the comb-box by the vacuum cleaner system.

The whole brushing station can be adjusted up and down to suit the tale top level by means of four adjustable pads. The two large orifices in the sideplates enable any damaged brushes to be removed and replaced quickly, without removing the sideplates.

Viewing Station Introduction The viewing station is the part of the machine where the measurements of reflected light are taken. It is this information which, when processed by the measurement computer, enables the colour of a pelt to be determined.

Design Configuration The viewing station (Figure 6.15) is made up of four major items: 1. A reciprocating carriage which houses the photodiode and filters used for colour evaluation. The carriage also has a vacuum head to prepare the pelt immediately before the measurements are taken.

2. A projector which casts a spot of light onto the part of the pelt which is being viewed.

3. A mask, against which the pelt is held during measurements.

4. A framework which carries the projector, mask and reciprocating carriage, and which is bolted to the main machine framework.

5. A mechanism ("Cardan drive") which generates the linear motion required for the carriage from the rotary motion of the main drive shafts of the machine. This mechanism is mounted directly on the machine framework.

Carriage Design The carriage is shown in Figure 6.16. When measurements are being taken it is moving from left to right.

Hence the first part to cross the pelt is the vacuum head. This consists of a nozzle attached to the front of the carriage which acts on the pelt through the slot in the mask. At the end of its stroke the nozzle is covered by the mask so that the pelt can drop away from the mask to be transported to the next station. At the trailing side of the vacuum head there is provided a knife edge which, together with the vacuum action, lifts the fur against the nap. The projector is aimed at the fur as it falls away from the knife edge. This can almost completely eliminate specular reflection from the surface of the fur, and leads to outstandingly accurate and consistent readings. Details of the vacuum head and knife edge are shown in Figures 6. and 6. . The optical part of the carriage is enclosed, except at the bottom, to prevent the ingress of dust which would distort the readings.The light from the projector enters through a glass window.

Only that part of the pelt actually being viewed is illuminated. This is achieved by means of a mirror, mounted at 45" to the axis of reciprocation, which reflects the beam of light from the projector onto the pelt immediately beneath it. Thus the patch of light moves along the pelt with the carriage. An aperture plate beneath the mirror ensures that the patch of light is of the size and shape required. A 2 mm diameter hole drilled in the centre of the mirror allows light to reach a photodiode placed behind the mirror. This photodiode provides the feedback for the projector lamp servo (section 7.9). This compensates for the varying distance from the projector to the carriage and for fluctuations in the mains supply, maintaining a constant level of illumination. An adjusting screw is provided to set the angle of the mirror during installation.The mirror will then be fixed in position with a dowel.

The light reflected from the pelt is measured by a second photodiode, which is mounted in a housing with an aperture at one end to limit the field of view. The optimum viewing angle was determined experimentally to beSS" to the vertical, and the optimum height 45 mm. It was thought necessary to provide a facility for viewing the pelt through up to three filters. These had to be used at the same station to reduce problems due to small differences in pelt preparation. The solution was to mount the filters on a disc which spins in front of the photodiode. To reduce the speed of rotation required, two filters of each colour are mounted on the disc.

Thus for one revolution of the disc six readings are taken, two with each colour filter. The speed of rotation of the disc is linked to the speed of the machine by driving the disc with a stepper motor using pulses generated by a digitiser on the machine drive shaft. (See section 6.2). Information about which filter, if any, is in front of the photodiode is decoded by logic circuits (see section 7.3.2) from the outputs of two HEDS 1000 photosensors which detect marks painted in the discs. The disc and photodiode are mounted in a sealed housing to exclude dust.

In practice, the pelts can be sorted entirely by intensity of light reflected, calculated as an average of the three readings. As an extra check on the working of the system, blue, yellow and white filters are used, and the white reading is compared with the sum of the other two. Some colour data may, however, be used to aid in the sorting. Since the machine has only to sort pelts consistently against essentially arbitrary criteria, it is never necessary to convert the output of the photodiode into an absolute brightness or reflectance; it can be handled throughout merely as a voltage.

The various components of the viewing carriage are mounted between two plates. At one end of the carriage there is the drive pin which links the carriage, through a con rod made of two rose bearings with a turnbuckle for adjustment, to the Cardan drive. On the outsides of the two plates, the housings for ball bushings are fitted, using shims to space them perfectly. It was decided to use ball bushings because they effectively restrain unwanted movements whilst allowing the long stroke of linear motion at the speeds which are required.

Projector The projector is used to illuminate a patch of the pelt. The projector shines along the axis of movement of the carriage, illuminating the pelt by reflection through a mirror. To maintain a reasonably uniform level of illumination throughout the stroke of the carriage, the projector has to give a parallel beam of light about 50 mm in diameter.

A quartz-iodide 250W 24V bulb is used. The light from this bulb is collected by an aspheric lens positioned so that the filament is at its focal point. This lens has an anti-reflection coating because of the high angles of incidence of the light rays. The second lens of the condenser is of much longer focal length (127 mm), it focuses the beam to a spot in the centre of an aperture plate. This plate is intended to cut out those rays which are too far off the optical axis to form a parallel beam when they pass through the objective lens which has the aperture at its focal point. The optical system is shown in Figure 6.17.

More than 95% of the output of the bulb is heat and to remove this, a heat filter is placed between the two lenses of the condenser. An axial flow fan mounted on the side of the projector blows cooling air through the condenser. This air is filtered to remove dust. A thermal switch is provided to warn of overheating due to fan failure.

Mask During viewing, the pelt is held against a mask by a pad which lifts it from the table (see section 6.4). The mask is fixed to the viewing station frame and ensures that the pelt is always held at the correct height for viewing.

The pad which presses the pelt against the mask is resilient and this, together with the half round bars running along either side of the mask aperture, ensures that the section of pelt visible through the aperture forms a smooth curve. This eliminates variations in readings, which could be caused by uneveness in the pelt.

The third purpose of the mask is to cover the vacuum head at the end of its stroke so that the pelt can drop away and be transported to the next station.

Cardan Drive The Cardan drive is a mechanism to provide linear simple harmonic motion to the carriage from the rotary motion of the drive shaft. It is used because of its compactness (the diameter of the body is less than the stroke of the output) and because of its ability to give a straight line output.

The principle of operation is shown in Figures 6.18 and 6.19.

The method of construction chosen was to use a series of plates dowelled together to make a body which completely enclosed the gears but which could still be assembled. The control shaft of the unit has the centre gear pinned to it and is used to mount the Cardan drive to the machine frame. The body of the unit acts as a cage carrying the three gears. The drive is transmitted to this cage through a timing belt pulley which is pinned to the underside of the body.

Distribution Module Design Requirements The Distribution Module has to segregate nine distinct shades of pelt plus any pelt that could not be classified. The Distribution Module accommodates three conveyors each divided lengthwise into three channnels, which take the graded pelts to one of three destinations. From this point they would be manually loaded into appropriate containers (see Figure 6.20).

General Module Configuration The distribution module configuration can best be split into two main area. Firstly, the walking beam mechanism (for further details see section 6.3) and secondly the distribution doors. The doors are mounted on a frame in three sets with one pitch separating each set. This was done to facilitate the positioning of the transportation conveyors under the doors. The unclassified pelts are progressed through the module and ejected at the end of the module.

The door frame is rigid in its own right and fits into the main frame which supports all the relevant parts in their correct configuration.

The trap-door does not fully open to allow the pelt to drop through as the original design did. Instead it drops down at the trailing end and the pelt is then pushed forward through the resulting gap by the walking beam mechanism (see Figures 6.22 and 6.23).

This requires a modified walking beam to be used in this module to ensure that the pelt remains under control until it has left the doors. It is necessary to make the rise and fall of the fingers a minimum of 65 mm.

This was achieved by using a lever to magnify the cam's rise and fall.

Door Design The doors allow the fingers to protrude through by 30 mm and still allow the pelt from the preceding door to pass underneath freely.

These requirements mean that underneath the leading edge of the door there has to be no incumbrance to stop the pelt, which in turn means that the only place where the door can be linked from side to side was at the trailing end. This results in a door with a triangular section.

The door has three thin top plates with gaps between them to allow the walking beam fingers to pass through the plates being rivetted to triangular supports.

Rivetting ensures that the drag on the pelts was minimised. The supports are joined by a bar which runs the full width of the door. This design (see Figure 6.24) allows the weight of the door to be kept to a minimum, but gives the structure the required rigidity.

The cost of the door could be halved by adopting casting for the more numerous and expensive items, such as the supports. The designs have been made suitable for casting.

Door Control Mechanism Design The design allows the doors to be synchronised with the walking beam by the use of cams. This is essential to the successful operation of the ejection system. If this had not been done, one of two faults would have occurred. First, if the door opened in advance of the fingers, the fingers could have pierced the pelt. Secondly, if the fingers retracted in advance of the door closing; the pelt, on the open door, could have moved forward, causing either the pelt to be ejected at the wrong station, or the door to trap the pelt.

The cam mechanism works, through a series of links, by pushing on an 'L' shaped lever to close the door.

At the end of each cycle the doors are closed and latched. When the cycle reaches the point at which the pelts are to be moved forward, the cam retracts allowing the doors to open in a controlled manner, but only the pre-selected doors are allowed to open. This is achieved by releasing the latch using a solenoid.

The levers are positioned at either end of the doors and supported on a frame which runs the full length of the module, (see Figure 6.26). The lever pivots concentrically with the door and is connected to it by the linking bar of the door. The levers are controlled by a bar which is in turn connected to the controlling cam by a one-to-one lever. This is shown in Figure 6.25.

Nose Rail The purpose of the nose rail is to provide a reference edge, when the pelt is loaded and a guide to keep the pelt at the centre of the table top as it is transported through the inspection module. The rail is made of 20 x 20 RHS and fastened to the upper frame via two telescopic arms. This allows adjustment to be made for different lengths of pelt.

Also on the nose rail are five brackets to accommodate sensors. These sensors are used by the machine control system to check that the pelts are in the correct positions.

Vacuum System The vacuum system of the machine performs four functions: 1. Clamping the pelts onto the table while it is brushed at the brushing station.

2. Removing dust and dirt from the brushes.

3. Performing the final cleaning and raising of the nap of the fur at the viewing station.

4. Cleaning down of the machine atthe end of a shift.

Selection of Vacuum Units The pelt clamping requires a hard vacuum but very little air flow. The other functions require a considerable air flow at near atmospheric pressures. These two combinations of conditions can both be obtained from the same vacuum unit but not simultaneously. Therefore it was decided to use two identical vacuum units.

The Z4 blower/extractor made by BVC Ltd. is used. It offers the following features of particular interest to us: 1. A.C. induction motor suitable for continuous operation.

2. Compact size and shape able to fit into the machine framework.

3. Output filter to B.S. 3928 removing 99.997% particles down to 0.6 microns - hence reduced health hazards.

4. Availability of a range of fittings, hoses etc.

5. Separately mounted filter drum with paper liners easy for cleaning.

The two units are mounted in the inspection module at opposite ends of the cradle. The filter drums project from the sides of the module for ease of access when changing bags. This arrangement also makes it easy to disconnect a unit so that it may be used as a conventional vacuum cleaner for cleaning the machine the end of a shift.

Clamp Value A pelt can be adequately held during brushing by extracting air through a pattern of holes under the pelt. It is necessary to apply and release the vacuum within 1/8 second if the machine design speed is to be achieved.

This meant having as short a stroke as possible so that the valve itself could operate fast. It also meant that the valve must be as close to the clamping point as possible (so that the smallest volume of air had to be moved).

These features were achieved by having a pattern of 2 mm diameter holes drilled in three plates. The upper plate is set into the table top, the lower plate is connected to the vacuum hose and the middle plate is a slider. In one position the holes in all three plates line up and the vacuum is applied. A movement of only 2 mm by the slider closes all the holes and releases the pelt. The slider is moved by a pair of solenoids (one for each direction of movement) one of which is always on.

The timing of the valve is controlled by the switch which provides current to the solenoids. This switch is operated by a cam on one of the machine drive shafts so that it is synchronised with the rest of the machine.

Electrical System The overall electrical system of the machine is shown in block schematic form in Figure 7 and Figure 8.

More detailed block schematic diagrams of components of the machine are as follows: Figure 9 - Stepper motor drive system and interface.

Figure 10-Walking beam position interrupt generation.

Figure 11 - Pelt sensing logic.

Figure 12 - Machine status monitoring.

Figure 13 - Filter disc position sensing.

Flow chart representations of the operations carried out by the machine are as follows: Figure 14 - Flow chart for measurement software.

Figure 15 - Flow chart for polling routine.

Among the components of the machine are the following : Sensors. These devices are used in the machine forthe following functions: i) tracking pelts through the machine ii) sensing maximum pelt 'height' iii) checking forjammed pelts in the trapdoors.

Modulated infra-red photosensors (by DI-ELL) are used.

In addition to the main sensors above, proximity switches and magnetic reed switches are used for other detection functions in the machine such as excess torque, optics-carriage movement, masking-pad stroke and guard position.

Main Drive System The following mechanisms and systems in the machine are driven from a prime mover through a torque limiter and in appropriate transmission.

i) Pelt transporting mechanism (walking beams for both the inspection and the distribution modules).

ii) Brushing station.

iii) Cardan drive for the viewing-station carriage.

iv) Masking-pad.

v) Door-closing mechanism for the distribution module.

The main drive output speed is fully adjustable over a range of 20-40 rpm. The system comprises a 1.1 kW, 3 phase, 41 5V ac induction motor, mechanical variator and gearbox. A brake is also fitted to the motor.

Power Supplies The power supplies for the machine are as follows: i) 24 DC supply for relays, proximity switches, solenoids and indicators. 8A at 24V is needed.

ii) 15VDC and 5VDCsupply for the analogue and logic circuits (respectively). Linear power supply units are suitable.

iii) 10-30VDC stabilised power supply for the lamp-servo circuit.

iv) l0VDC (maximum) for the stepper motor system. An integral power supplystepper drive by "Digiplan" is used.

v) 15VDCsupply for the pelt sensors and associated CMOS interface electronics circuit. A power supply unit rated at 2A is used.

The computers and peripherals have their own built-in power supplies.

The machine is equipped with a set of three phase contactors to drive the conveyers and main motor.

Cooling fans are installed in the electrical/electronics cabinets (and the viewing station) to circulate air and deduce localised overheating. 1 0VAC and 1 2VAC transformers are provided for the contactors and earth protection relays respectively, and a 1 3A socket strip is also provided for use of test equipment and instruments during maintenance and repair. The power drawn from each phase of the mains supply is approximately balanced by distributing the loads as required.

a) Initialisation. With the main isolator switched on, power is supplied to the computers, cooling fans, power supplies and transformers. A "keep alive" relay, energised by the control computer, enables normal machine operation and disables the machine in the event of computer malfunction. The keyswitch (SW1 ) allows either normal machine operation or exclusive access to the computer keyboard for changing pelt input perameters. A yellow light (standby) on the display panel illuminates when the computer reports a safe start-up condition.

b) Start-Up. When the 'start' button is pressed the automatic start cycle will be activated and the yellow light will go out providing: i) All protective guards are closed.

ii) The earth protective relays are made.

iii) The emergency stop buttons are in the 'off' position.

The computer will then start the following: iv) The main motor.

v) The vacuum units.

vi) The lamp servo and stepper motor systems.

vii) The conveyor drives.

A green light will be displayed on the panel when all systems are functioning satisfactorily, indicating to the operator that pelt loading can commence.

c) Normal Halt. Depressing the belt button will cause a red light to come on and the green to go off. The computer will then stop the machine in two stages: i) Motor shut down.

ii) Shut down of remaining systems.

When a complete shut down has been achieved, the yellow light will again be displayed and the red will go off. This indicates to the operator that the machine can now either be switched off (at the isolator switch) or re-started.

d) Emergency Stop. An emergency stop can either be computer initiated due to machine malfunction or other situations (such as an inadvertent opening of a guarded cover during machine operation) or manually initiated by depressing any of 4 emergency stop switches located strategically around the machine. A total machine shut down results (with the exception of the fans and computers) and an emergency stop light is displayed on the panel. The emergency stop switches are reset manually.

TWO high resolution optical reflective sensors (Hewlett Packard HEDS 1000) are used to obtain: a) a pulse for each filter which is fed into a three bit counter, b) a reference pulse used as a datum and to reset the counter.

These sensors are fully integrated modules designed for optical reflective sensing. Each module contains visible LED emitter and matched IC Photodetector. A bifurcated aspheric lens is used to image the active areas of the emitter and the detector to a single spot 4.27 mm in front of the package.

The sensors are interfaced with the LSTTL circuitry in directly coupled configuration, and logic level chatter reduced using Schmitt triggers.

The circuit has flexibility which means that data can be read through three sets of two filters instead of two sets of three filters by simply modifying a few connections.

Analogue to Digital Converter (A DC) This subsystem, due to the requirements of speed and high precision and in order to allow for different speeds of rotation of the filter disc, is a modified Dual Slop ADC. Instead of having a reference voltage applied to the integrator during the discharge period, the output from another integrator (charged from a constant voltage reference source) is applied. The reason for this modification is that any variation in the conversion time must not affect the value of the readings.

The operation of the ADC is as follows: a) Charge Up. While a filter is in front of the photodiode the analogue signal from the viewing head measuring amplifier is fed to an Integrator (11). During the same period, another Integrator (12) is charged from a very precise reference voltage.

b) Discharge. In this period, both in-coming signals (photodiode and reference voltage) are switched to earth and the output from the reference integrator (12) is used to discharge the signal integrator (I 1). When this process starts, a twelve bit digital counter is started by the control logic and counts until the discharge is complete. This count represents the digital value of the reading taken during the charge up speed.

c) Auto Zero. This cycle is used to compensate for any voltage drift at the inputs of the Integrators. The zero crossing detector (comparator) is used as an amplifier and switched to close a unity gain loop round the integrator. The integrator input is earthed and the comparator charges the auto zero capacitors (CO and CO) with a voltage equal to the offset, which is then subtracted from the input of the integratorfor the rest of the cycle.

External Memory The twelve bit value stored in the counter of the ADC corresponds to one reading from a pelt. A set of readings is needed for the Measuring Computer to decide on the shade of the pelt. These values are stored in an external Random Access Memory (RAM) until they are required by the computer.

When a signal coming from the filter disc position sensor is detected by the Control Logic the process starts and the values stored in the ADC counter are transferred to the RAM at the end of each count. Signals from the Address Generation circuit are used by the control logic to generate three different sets of subsequent addresses, corresponding to the three different sets of filters in the filter disc. When all the readings have been taken (the number of readings may be adjusted using a dual in line switch on the circuit board) the internal Address Generation Logic is disabled and an Interrupt signal is sent to the measurement computer interface (GPIO), this unit sends back a control signal and starts sending 7 bit Addresses to the Control Logic, these addresses correspond to the RAM locations in which the Data is held.

Every time the computer generates an address, the Data held there appears at the output of the RAM and the computer reads this value from the GPIO. This method of Interface with the computer is known as Partial Handshake. This process occurs while pelts are being transferred between the machine stations.

HardwiredAverage The clock pulses fed to the ADC counter are also fed into the Hardwired Average circuit. This consists of three similar subcircuits one corresponding to each of the three filters.

The control logic select averager circuit establishes from signals provided by the filter disc position circuit which filter corresponds to the data being recorded and sends the clock pulses to the selected subcircuit. In each subcircuit, the clock pulses are sent to a divide by N counter (the value of N being that of the number of readings per filter) and the output of this unit is sent to a 12 bit counter which will hold the average value

~ ~ Digital Data i=1 N Multiplexer When a full set of readings has been taken the following Data is held in the system:: a) 3 x N 12 bitword in the RAM b) Average 1, a 12 bit word corresponding to filter 1 c) Average 2, a 12 bit word corresponding to filter 2 d) Average 3, a 12 bit word corresponding to filter 3 Since the measurement computer has to have access to all this data through 12 bits of the GPIO, it has to be multiplexed. The multiplexer control logic uses information from the 7 bit address generated by the measurement computer. This allows the data from either a, b, cord, to be transferred.

Filter Disc Drive System Stepper Motor The machine measures light through 2 sets of 3 filters mounted in the filter disc.

The maximum required speed for the rotation of the filter disc is 1440 rpm. This rotation of the filter drive is achieved by a stepper motor as opposed to conventional AC or DC motors, because the stepper motor offers the following advantages: i) A stepper motor is inherently a discrete motion device. It is more compatible with modern digital control techniques. The stepper motor is easily adapted for interfacing with other digital components.

ii) The positional error in a step motor is non-cumulative.

iii) It is possible to achieve accurate position and speed control with a step motor in an open loop system, thus avoiding ordinary instability problems.

iv) Design procedure is easier for a step motor control system.

A SIGMA permanent magnet rotor type with step angle 1.8 is used.

Stepper Motor Drive (Figure 3) In order to drive the stepper motor, a high performance stepper motor drive (Digiplan type 1054) is used. It has the following advantages: i) Bipolar, bilevel drive circuit combing smooth performance with efficient delivery of up to 200 watts of motor shaft power. The bipolar system gives the highesttorque-per-wattfrom the motor. Since all the windings are fully utilised - current is always flowing in one direction or the other in each winding. The use of bilevel switching is largely responsible for the high overall efficiency and there is a complete absence of the audible noise associated with chopper - regulated drives at standstill. The system uses both high and low voltage power supplies, the high voltage supply being used to overcome winding inductance and ensure a rapid build-up of motor current. When the current reaches the required level the high voltage supply is switched off and the low voltage supply takes over, minimising the drive circuit losses. The current is controlled by an active regulator which acts as a powerful damper on motor resonance effects.

ii) Either full or half step operations may be selected, giving 200 or 400 steps/rev with the 200 steps/rev motor. The torque-speed characteristic may be optimised for the task by adjustment of current levels and it is even possible to change the drive characteristics during operation by using the boost facility. This can be very useful when the load on the motor changes during its operating cycle.

iii) In addition to the usual step and direction controls, the drive has a de-energise input which allows the motors to be turned by hand. The boost control enables motor power to be increased during periods when extra torque is required and so avoids the need to run continuously at maximum power. Motor current is automatically reduced at standstill to a presettable level.

Drive Interface (Figure 3) The main components of the drive system include the translator, switch set and power supply. The drive is interfaced with a logic circuit which extracts the direction of rotation of the main motor by means of a digitiser, which sends two pulses inphase and quadrature for each increment of the shaft rotation.

The basic input signal required by the system is a series of uniform clock pulses, the number of pulses corresponding to the number of steps the motor is required to perform. From this pulse sequence are generated a set of controlled drive waveforms which, when applied to the stepping motor, cause it to execute the required number of steps.

The first part of the process is carried out by the translator It generates a series of low-level signals which is fed to the switch set to control both timing and direction of currents flowing in the motor windings. In generating these signals the translator takes account of the direction pulses supplied by the logic circuit.

Digitizer and its Interface Design An optical incremental shaft encoder (i.e. a digitizer) is used for measuring shaft position. Incremental techniques are also used for measuring speed etc.

The digitizer in the control system is required to: i) determine the incremental position of main shaft ii) synchronize the speed of the stepper motor, and thus the rotating filter disc, with that of the main motor (at maximum speed of the main motor (40 rpm) the stepper motor should be running at 1440 rpm).

iii) Generate interrupts at various positions of the walking beam.

iv) Control the hardwired averager.

The digitizer has a resolution of 1800 lines with the added facility of a zero marker. It consists of a glass disc with 1800 lines accurately generated at regular intervals. These lines are used to 'make and break' an optical switch, and the resultant signal is amplified to give a square wave output. Since this output is a pulse for every increment of rotation, external circuitry is required to integrate the pulse train and indicate the shaft's position.

The digitizer is manufactured with two sets of lines which are phase shifted by 90". This allows the direction of rotation of the shaft to be determined by adding an additional logic circuit. This automatically gives a count pulse rate offourtimes the number of lines. Thus the resolution was increased to 7200 lines (4 x 1800). This circuit also contains up/down counters which count pulses from the direction logic according to the direction of rotations. The zeromarker pulse from the digitizer is latched and used to reset the counter to zero. The circuit also avoids the multiple count problem which can occur in the event of a stationary shaft oscillating slightly about a transition.The digital outputs of the digitizer were also buffered by LSTTL Schmitt triggers which increased the fan out capabilities and lowered the system's susceptibility to errors caused by slow transition times of the digitizer's outputs.

Walking Beam Interrupt Generation - (Figure 4) The control computer is interrupted six times per machine cycle to perform various functions at certain positions of the walking beam (e.g. Open the trapdoors, check the machine status word etc.). The software method of checking the digitizer pulses would be time consuming. Therefore, the hardwired method is used.

The principle of this circuit is shown. An eight bit position word is sent to a comparator, from the control computer via its interfaces. When the walking beam reaches the required position, the comparator sends a pulse to the control computer and an interrupt is generated.

Pelt for Minute Counter The purpose of the Pelt Per Minute Counter (PPMC) is to determine the number of machine cycles per minute and to compare this number with a target value, set by the machine supervisor (e.g. 40 rpm). Should the value, as shown on a 31/2 digit liquid crystal display, be over or under the pre-set value, the machine will stop automatically, indicating that a fault condition exists.

The PPMC is an automatic re-cycling counter and so regularly updates the count information. To accomplish this the following sequences are: a) The counter gate is opened for precisely 0.Ssec. to allow the input pulses from the encoder to be counted.

b) The information from the completed count is transferred to the bistable latch circuit and displayed.

c) The latched information is transferred to the comparator circuit in which it is compared with the preset values.

d) The counter is then reset to zero (cleared) and the gate start and stop controls are reactivated to enable the next measurement cycle to begin.

An internal oscillator generates pulses to operate the gate, the latch and the counter reset. The internal sequence begins when a rising edge from the gate generator occurs at the start input of the counting gate.

This causes the gate to open to count the pulses from the digitizer circuit. The counting interval ends when a low to high transition appears at the stop inout. This also triggers the latch pulse generator. The latch pulse, a momentary low of a few micro seconds, causes the accumulated count to be stored and displayed. When the latch pulse returns to high the counter can be reset without affecting the contents of the latch. The rising edge of the latch pulse triggers a short delay generator which in turn, triggers the counter reset-pulse generator. When the counter is reset, the gate generator is reactivated so that another cycle can start.

At the same time the latched count is transferred to the comparator. There it is compared with the value entered by the operator. In the event of a discrepancy, a signal is sent to the control computer to store the machine.

Pelt Sensing Logic (Figure 5) There are 26 sensors in the machine including pelt tracking sensors, door closure sensors, etc.

The 15V output signals from the sensors are first filtered using low pass filters following by buffers to increase the current drive capabilities of the signal lines.

These signals are then latched before going into the control computer through the multiplexers. The latches are necessary to retain the information for some period of time until the computer can read through the multiplexer and data bus. When the data transfer is completed, the computer resets all the latches.

The Lamp Servo Control The light required for illuminating the pelts in the viewing station must be of constant intensity irrespective of lamp ageing position of the viewing head, mains variations, etc. Conventional power supplies are not suitable since they are unable to compensate for such factors as lamp ageing and the varying distance between the lamp and the pelt.

The lamp servo is a module designed to provide constant illumination in the automatic colour evaluator.

The servo includes a Programmable Power Supply (PPS) and optical feedback and control circuits. The optical feedback circuit makes use of a photodiode (type BPX65 (RS 304-346)) which is incorporated into the viewing station carriage to sense the intensity of the beam from the projector lamp.

The overall system is designed to act as a closed loop servo.

The lamp servo is started by a signal from the control computer.

A 'soft-start' circuit is provided in the lamp servo design. The soft-start circuit ensures that the lamp current is gradually built up to its normal operating value. This feature is necessary because the tungsten filament in the lamp has a very low resistance when it is cold and the current drawn could exceed the capacity of the power supply if it were switched on directly.

An important advantage of the soft-start feature is that lamp life will be enhanced. The lamp is of a type used in many slide projectors and has a nominal rating of 250 watts at 24V.

Photodiode and Pre-Amplifier Circuit Design The photodiode is operated in a linear reverse biased mode such that it acts as a current source to arc inverting pre-amplifier stage. The pre-amplifier is based on an LM308 op amp, additional components are used to ensure low noise and drift. A trimming potentiometer is used to allow for the 'dark' current of the photodiode. The pre-amp is mounted on the carriage of the viewing station away from the main lamp servo control circuit.

The PCB Lamp Servo Circuit The PCB Lamp Servo Circuit incorporates the following individual circuits a) The TTL input circuit b) The soft start circuit c) The system gain amplifier d) The error amplifier e) The Lamp Servo Failure Detector f) The Temperature Sensor Circuit The TTL input circuit makes use of an SN72741 op-amp connected as a simple comparator and its output is normally at OV. On receipt of the start signal the output rises to + 1 SV and this signal is applied to the soft start integrator.

The soft-start circuit makes use of an LF353B dual op-amp. One section of which is used in a integrator circuit to charge a capacitor at a linear rate. A diode, ensures that the integrator output is clamped at approximately zero volts before the integrator is initiated by the TTL input signal. The output ramp signal from the integrator is applied to the other section of the LF353B which is connected as a comparator circuit in which a diode clamps the ramp signal to a reference voltage of 10 volts. The reference voltage is externally supplied to the circuit from the ADC (Analogue-to-Digital Converter). The output signal of the soft-start circuit therefore provides a delay in the lamp servo control which allows the current in the lamp to build up gradually.

The system gain and error-amplifier circuits use the LM353B dual op-amp. The input signals for the system gain amplifier come from the photodiode via the pre-amp. The design of the system - gain amplifier includes trimming pots which are adjusted to provide the system with sufficient gain.

The error-amplifier has opposite phase input from both the system gain amplifier and the soft-start circuitry. Functionally, the error-am p adjusts the controlling voltage to the programmable powersu-pplyfor - the lamp servo system. A continuous adjustment is achieved by the use of feed back inputs provided by-the photodiode. A 5.6V Zener diode limits the outout of the error-amplifier.

The servo PCB has a logic output signal which is interfaced to the computer, the purpose of this is to ensure that the lamp servo is continuously monitored. Any malfunction detected in the lamp servo will cause the circuit to provide a logic one output which causes the computer to shut down the machine system.

The temperature near the bulb is monitored by a temperature sensor. The sensor is connected to its circuitry on the lamp servo PCB. Any overheating detected by the temperature sensor will cause the circuit to generate a logic one output causing the computer to shut down the machine.

Both logic outputs for the computer to disenable the machine system are logically combined in OR gate and the resultant signal is one bit of the machine status word.

Programmable Power Supply (PPS) This special form of power supply unit has facilities to adjust and set the output voltage or current to a known value by means of an external programmable input which, in this case, takes the form of a voltage signal. The lamp servo supplies the voltage control signal. The configuration adopted uses the PPS as the main power-amplifierforthe system. The PPS has a nominal output of 33V maximum at 12 Amps.

The Computer System The computing system for the machine satisfies the following constraints: i) The computing speed enables the machine to process 40 pelts per minute.

ii) A high level programming language, suitable for machine control, is used.

The system includes two Hewlett Packard 9915A microcomputers programed in BASIC. The machine function is divided, into two parts: the pelt colour evaluation and the machine sequencing control, and the two units are arranged as measurement and control computers, respectively. Both units are approximately equal amounts of real time to execute their respective functions and only one piece of information, the pelt shade need be passed between the computers during the machine cycle.

Before the start of each run of the machine, information regarding the pelts to be processed (e.g. size, sex and colour type - dark, pastel, etc.) are put into the computing system from the keyboard, located in the control cabinet. This enables the two computers to open (identical) data files on their respective cassettes, and write appropriate headings to them. In addition to this the Measurement computer reads from its cassette the required colour chart, which provides a reference tablet for a decision to be made regarding colour shade. Located immediately above the keyboard is the visual display unit which serves not only to display information being put into the computer but also to indicate the nature of any faults which occur during machine operation.

Data concerning the numbers of pelts processed in each shade category are written to the data files on both cassettes when the machine is halted for ariy reason. To protect this data in the event of machine power failure the units are equipped with battery power supplies. These will maintain computer operation for approximately 15 minutes which should be ample to enable the necessary cassette storage.

The Measurement Software - (Figure 8) During every machine cycle the measurement computer receives light intensity readings taken from the viewing head, ànd also averages of those readings. For many pelts the averages so obtained would be sufficient to enable the measurement computer, after consultation of its colour chart, to indicate to the control computer the shade of the pelt concerned. However, the measurement computer does more than this. In order to counteract effects of any marked changes in colour along the pelt (caused by white spots, for example) each reading is compared with the average of the reading set.If the reading differs from the average by more than a certain percentage, then the reading is noted and after all readings in the group have been scrutinised, the average is re-calculated, excluding the effects of any non-conformed readings. This value is compared with the colour chart to decide the pelt shade. This slide is then held in memory until requested by the control computer.

In the event of too many of the readings failing the "acceptability criterion" the computer recognises that it no longer has sufficient data to make a valid judgement and automatically grades the pelt into the "reject" category.

Sequencing Software The software controlling the Automatic Colour Evaluator can be divided into two major types of program, which are: a) Programs that are executed with 'real time' constraints.

b) Programs fhat do not require 'immediate' actions arid are therefore independent of 'real time'.

The real time operations are those tasks that must be executed and completed within each cycle of the machine and consequently their total duration must be less than the minimum cycle time for the machine.

The other programs control tasks that have no special constraints on their execution times i.e. the software operator at its own speed waiting indefinitely, if necessary, for inputs.

All the software on the automatic colour evaluator is written in an extended version of BASIC specially structured for the control of machines and processes. The sequencing software is implemented in a high level language for the following reasons: i) The processes within the machine are not unduly complex and are fully interlocked mechanically.

ii) The development of programs is simpler and less prone to error.

iii) The maintenance and any subsequent modifications of the software by the user is simplified.

The sequencing software is executed by a Hewlett Packard 9915A microcomputer to which various standard interfaces and ROMs have been added to achieve the designed input/output structure and to aid program development. The storage medium for all the software is cassette tape cartridges although the 9915A does offerthe option of using EPROMs as an alternative.

Real Time Software The three major operations performed by the real time software are: a) The monitoring of the different machines states that occur during each machine cycle.

b) Opening of the correct door in the distribution module for any particular pelt according to the evaluation process previously performed on that pelt by the viewing station and the measurement computer.

c) 'Tracking' all the pelts which have been loaded as they travel the length of the machine.

Other tasks carried out by the real time software include counting the number of pelts of each shade that have been processed, and monitoring the number of pelts that have been rejected for one reason or another.

The computer also monitors its own performance and if any discrepancies are detected by this process, or any other of the monitoring routines, the machine will shut down in a controlled sequence which is independent of any software.

The execution of the real time software is controlled by a series of interrupts which direct the program to various polling routines and hence perform the tasks in a), b) and c) above. The interrupts are created at different stages of the machine cycle by an output signal from the comparator in the digitizer circuitry or by the internal timers of 9915 itself. Atypical example of a polling routine which monitors a specific state of the machine is illustrated by the flow chart in Figure In addition to the events occurring within the cycle of the machine the sequencing software must also monitor some external hardware i.e. the control buttons on the operator's panel, temperature sensors etc.

This task is performed by the microprocessor when it has been initialised by the time independent software.

An external event detected by this microprocessor, or 'Watchdog', causes the real time software to take the appropriate actions.

Time Independent Software The Time Independent Software is concerned with three main procedures which are: a) Operator Intervention The operator, if he wishes, can change the door number that is allocated to any particular shade and modify the acceptable reject rate above which the machine will shut down. These two operations can only be performed when the machine is stopped and access to the control has been gained. If the above parameters are not modified then default values are used.

b) Machine Monitoring The Time Independent Software monitors the state of the machine before it starts the main motor and executes any Real Time Software. During this monitoring period the shade counters and other variables are changed to allow for pelts that have been removed from the machine due to unexpected jamming or other faults.

c) Storage of Data At the end of each batch of pelts or if a mains power failure occurs, the contents of the shade counters, the number of pelts processed etc. are written to the cassettes on both the control computer and the measurement computer. This is done as a form of data integrity and allows the retrieved and analysis of the data by other similar computers i.e. HP-85.

Control Cabinet The function of this item is to provide a station at which communication may take place between the operator and the machine. Information relating to the status of the machine is given by the following: i) The computer screen (or VDU) displays messages from the computer, relating to diagnostics of any errors which may have occurred. This is also used in conjunction with the computer keyboard.

ii) Other diagnostics are supplied by the "tell tale" lights which are normally illuminated and an error or failure state is indicated by a particular light switching off. These give information relating to temperature on the electrical cabinets, to whether any guards have not been properly closed and to whether any emergency stop switches have been left depressed.

iii) The status of the machine function is indicated on the left of the VDU. The three lamps at the top show from left to right, whether the machine is stopped, but switched on or if the machine is in the warm-up state or finally whether loading of the machine may commence or continue. Below these lights a counter is situated which indicates the machine speed.

Communication of information from the operator to the computer is carried out using the keyboard which may be brought out of the cabinet or extendable slides. This piece of equipment which may be used whilst the machine is not operating, is enabled by means of the key-switch situated immediately above the keyboard and to the left of the VDU. On the same level as this, to the right of the screen, is provided a computer port for a printer, thus facilitating the production of "hard copies" of information stored in the computer.

Claims (33)

1. Apparatus for optical measurement of a surface, comprising means for applying suction to a portion of the surface, and means for measuring the reflectance of a portion of the surface that has just been subjected to the suction.

2. Apparatus for optical measurement of a surface comprising means for ruffling a surface having a pile or nap, and means for measuring the reflectance of the ruffled portion of the surface.

3. Apparatus as claimed in claim 2, or in both claim 1 and claim 2, wherein the ruffling means comprises a straight edge and/or the suction means of claim 1.

4. Apparatus as claimed in any one of claims 1 to 3, wherein the measuring means comprises means for projecting a beam of light onto the surface and means for measuring the brightness of the surface illuminated thereby.

5. Apparatus as claimed in any preceding claim, wherein the suction of ruffling means and the point of action of the measuring means are arranged to be moved together over the surface.

6. Apparatus as claimed in any one of claims 1 to 5, wherein in operation the surface under measurement is stationary, and which comprises a stationary source of collimated light parallel to the surface, which is caused to scan across the surface by being deflected by a moving, angled, mirror.

7. Apparatus as claimed in claim 6 wherein the intensity of the light is measured at the mirror and is maintained constant by a servochanism.

8. Apparatus as claimed in claim 4 or claim 6 or claim 7, wherein the collimated light falls on the surface substantially perpendicularly.

9. Apparatus as claimed in any one of claims 1 to 8, wherein the measuring means views the surface at an oblique angle.

10. Apparatus as claimed in claim 9, wherein the measuring means faces in the direction from which the surface emerges from under the suction or ruffling means.

11. Apparatus as claimed in any one of claims 1 to 19, which comprises an apertured mask means for holding a surface to be measured against the mask, and means for causing the measuring means and the suction or ruffling means to scan across the aperture.

12. Apparatus as claimed in claim 11, wherein the scanning means comprises a reciprocating carriage.

13. Apparatus as claimed in claim 11 or claim 12, wherein the holding means comprises a resilient pad and means for pressing the pad against the mask.

14. Apparatus for brushing, comprising an endless belt bearing brushes over a minor portion of its length, means for causing the belt to move continuously along its length, and means for causing articles to be brushed to move intermittently past the belt in such a manner that each such article is held stationary at a point on the path of the brushes while they pass, and another article is substituted between passes of the brushes at that point.

15. Vacuum holding apparatus comprising a first perforated sheet to which articles are to be held, a second sheet parallel to and spaced apart from the first sheet, and having holes in register with those in the first sheet, or third perforated sheet occupying the space between the first and second sheets and movable between a position in which its perforations are in register with those in the first and second sheets and a position in which it closes off those perforations, and means for applying a vacuum to the outside of the second sheet.

16. Apparatus for sorting articles, comprising a row of trapdoors, means for advancing articles onto each trapdoor in turn along the row, means for selectively opening said trapdoors to permit articles to pass through them, and means for closing all the trapdoors.

17. Apparatus as claimed in claim 16, wherein each trapdoor, when open, inclines downwards to the trailing edge, wherein articles are advanced by fingers of or walking beam conveyor having a pitch equal to the spacing of adjacent trapdoors, and wherein the arrangement is such that an article on a door when the door opens is advanced off the edge of the door by the door by the conveyor.

18. Apparatus as claimed in claim 16 or claim 17, wherein the fingers of the conveyor are brought into engagement with the articles before the trapdoors are opened, and wherein the trapdoors are closed before the fingers are retracted to prevent an article that has been advanced onto an already open trapdoor from falling off the edge of the trapdoor.

19. Apparatus for classifying articles including means (third means) capable of measuring a characteristic of the articles means (fourth means) capable of storing a plurality of the said measurements, and means (fifth means) capable of (i) establishing an initial average value for the measurements, (ii) comparing each measurement with the initial average value, (iii) discarding measurements differing from the initial average by more than a set amount, (iv) establishing a revised average based on accepted measurements, (v) comparing the revised average with an average resident in the apparatus, and, (vi) signalling how the revised average value compares with the resident average value.

20. Apparatus as claimed in claim 20, wherein the fifth means is capable of rejecting all the measurements if the number of accepted measurements is below a set minimum.

21. Apparatus as claimed in claim 19 or claim 20, wherein the fifth means is capable of comparing the revised average value with a plurality of resident average values or ranges and of signalling which resident average value the revised value comes closest to, or which range it falls within.

22. Apparatus as claimed in claim 21, and incuding means capable of recording the number of times each resident average value is signalled by the fifth means.

23. Apparatus as claimed in any one of claims 1 to 13 and in any one of claims 19 to 22, wherein the said measuring means is the said third means.

24. Apparatus as claimed in any one of claims 16 to 18 and in claim 21 or claim 22, wherein the signal determines which trapdoor (if any) opens under the article in question.

25. Apparatus as claimed in any two or more of claims, 1, 14, 15, 16 and 19, and/or in any claims dependent upon the respective ones of those claims.

26. Apparatus as claimed in any preceding claim for assessing a reflectance characteristic of a surface having a nap or pile.

27. Apparatus as claimed in claim 26 for use on fur.

28. Apparatus as claimed in claim 27 for determining the shade of mink pelts.

29. Apparatus as claimed in any preceding claim and substantially as hereinbefore described with reference to any one or more of the drawings.

30. Apparatus as claimed in any preceding claim and substantially as hereinbefore described with reference to, and as shown in, any one or more of the accompanying drawings.

31. A method using apparatus as claimed in any preceding claim.

32. A plurality of mink or other pelts or other articles sorted by apparatus or a method as previously claimed as having substantially the same shade or other characteristic.

33. Any new or novel feature hereinbefore described or disclosed.