FI72816C - ICKE-FOERSTOERANDE FOERFARANDE OCH ANORDNING FOER INSPEKTION AV DAECK. - Google Patents

Description

1 72816

Non-destructive test method and apparatus for inspecting tires

This invention relates generally to a test piece non-destructive method and apparatus for inspecting rubber rings. Such an inspection technique may also be combined with conventional grinding operations in accordance with the present invention.

There has long been a strong need to provide a cost-effective, efficient, non-destructive test (NDI) test for rubber outer tires. There are obvious safety benefits to be gained from such technology if it is to be implemented efficiently and quickly. There are also potential economic benefits. For example, in tire retreading operations, a defective tire carcass can be discarded before more money and time is spent if the defective individual can be accurately, efficiently, and quickly detected.

In fact, the need for an improved NDI method and equipment for testing outer tires is so great that the U.S. Army Department, U.S. Pat. Special symposia have been published by the Army Materials and Mechanics Research Center to cover this area alone in 1973, 1974, 1976, and 1978. Documents for the first three symposia have now been published and are available from U.S. Pat. National Technical Information Service. Each includes a complete chapter on ultrasound testing of the tire, as well as other chapters dealing with different testing methods (e.g., holographic, infrared, X-ray). There are also many prior patents relating generally to the use of ultrasonic waves for non-destructive testing of a test piece on pneumatic rubber rings. Examples of these patents are U.S. Patents 2,345,679, 2,378,237, 3,336,794, 3,604,249, 3,815,407, 3,882,717 and 4,059,989.

There are also several prior patents relating to mechanical structures for attaching or otherwise physically treating pneumatic outer rings during various 2,72816 types of testing or manufacturing processes. Examples are U.S. Patents 2,695,520, 3,550,443, 3,948,094 and 4,023,407.

Although a very large number of different non-destructive ultrasonic tests have been performed on tires in the past in accordance with these prior patents, they all have shortcomings and have not been able to achieve widespread acceptance in commercial practice. Some of the prior art methods have required a liquid coupling agent on one or both sides of the wall of the test tire. Some previously known testing methods use the so-called a "pulse echo" procedure that produces a rather complex echo formation due to normal ring internal structures as well as abnormal structures. Many methods have used relatively low frequencies (e.g. 25 kHz), resulting in severe interference from normal ambient acoustic sources, while other methods have used extremely high frequencies (e.g.

2 MHz), resulting in rapid signal attenuation. Some methods have used continuous ultrasonic waves, resulting in a blurred echo pattern or standing waves and similar problems, while others have considered envelope peaks in received acoustic waves. Some methods have used individual pulses of the acoustic signal for each ring test site. In some cases, the magnitude of the received envelope peak has been used to reach the final numerical values. Attempts have also been made in some cases to test a fully inflated ring body (but sometimes causing acoustic signals to pass through two ring walls so that all transducers can be kept outside the ring), although in most cases attempts have been made to test empty ring bodies. There may also be techniques other than those mentioned above.

Prior art tire mounting mechanisms have generally included axially movable tire mounting rings for rapid mounting and inflation of the tire under test. In previous NDI machines, an ultrasonic transmitter has been placed inside a rotatable inflated ring, although these have been

II

3 7281 6 fixed or manual installation arrangements only. Other ND1 machines have included an articulated transmitter mounting arrangement with an unfolded, unfilled test tire. Until now, however, there has been no viable mechanical arrangement for the rapid placement of ultrasonic transducers along the wall of a inflated test ring, which at the same time would facilitate rapid tire attachment / removal operations and also protect transducers from physical inconvenience and damage.

It has now been found that these prior attempts to ultrasonically inspect outer tires without destroying a specimen can be greatly improved and made more commercially viable.

According to the invention, there is provided an apparatus for grinding a tire carcass and testing a ground carcass without breaking it, the apparatus being characterized by opposite circular mounting rings tightly attached to respective edges of said tire, ring filling means for filling said ring to rotate said inflated tire and grind the outer tread away to provide a substantially flat tread to the resulting tire carcass, ultrasonic testing means located both inside and outside the inflated ring to measure the relative intensities of the acoustic ultrasonic signals passing through the wear regions of said polished tire; the ring is rotated to test the ground of the ground tire without breaking it.

There is also provided a method of grinding a tire carcass and testing a carcassed tire carcass without breaking it, which method is characterized in that the method uses a single unitary device which secures the edges of said tire tightly between opposite retaining rings, fills said tire when said retaining rings are attached rotating said inflated tire and grinding the outer surfaces of the tread wall to obtain a substantially flat outer surface of the tread wall in the resulting tire carcass and measuring the relative intensities of the acoustic ultrasonic signals passing through the tread breaking.

According to one embodiment of the invention, the pulse or burst transmission mode can be used to reduce standing waves or unwanted reverberation effects within the ring. Each burst includes only a few (e.g., 100) cycles of the acoustic signal, resulting in a very short overall duty cycle and highly efficient transducer operation. At the same time, it has been found that the envelope of received acoustic signals can be altered by internal reverberation (reflections), standing wave, wave reversal or other irrelevant wave effects after the beginning of each burst, i.e. the rising edge. Accordingly, in a preferred embodiment of the present invention, the received acoustic signals are passed through a reception gate circuit so that only those signals from the burst that are at the beginning of the burst are utilized.

In some cases, further improvements are possible by averaging the values obtained at different frequencies, thus avoiding some of the potential adverse effects of the standing waveform and 5 7281 6 such adverse effects. In addition, a nonlinear analog-to-digital conversion technique can be used to promote usable values.

In the presently preferred embodiment, a plurality of transmitting acoustic transducers are placed inside a rotating air-inflated ring to acoustically "illuminate" the entire interior of the ring under test. However, it has been found that special wave rehabilitations, standing waveforms, or similar wave effects can be generated to distort readings if more than one transmitter is activated at a given time. Accordingly, a preferred embodiment includes a multiplexing circuit to ensure that only one transducer becomes activated at a given time.

A plurality of acoustic receiving transducers are disposed along the outer wall of the ring to receive acoustic signals transmitted through the wall from the transmitting transducers located inside the ring.

Each of the receiving transducers is preferably repaired and adapted to the impedance of the ambient air by means of a cylindrical tube with a conical inner surface which tapers towards the tactile surface of the actual receiving transducer. Such collimation helps to limit the output of each receiver to represent those acoustic signals that pass through a limited area of the tire wall and further helps to eliminate interference due to tread shape and environmental noise. Defects in the tire, such as detachment of cord layers, detachment of cord layers and rubber layers, and detachment of different rubber layers, attenuate acoustic signals passing through these layers more than what attenuates when they pass through a normal part of the tire wall.

It has also been found that leaks in a pressurized ring (i.e., air passage through the ring wall) can be detected with the same ultrasonic transducers by noticing an increase in the received signal level above the level obtained when the acoustic signal passes through the normal ring wall, even when the ultrasonic transmitter is turned off. .

Each of the receiving transducers is connected to its own signal processing channel, although multiple receivers may also be multiplexed to share a common signal processing channel in synchronism with the multiplexing of a plurality of acoustic transmitters so as to minimize the number of signal processing paths required. A relatively long-term automatically amplified amplifier is included in each signal processing channel to compensate for different average signal levels from one ring to another and from one channel to another depending on differences in the average wall thicknesses of the rings. After the automatically gain-adjusted gain, the received ultrasonic signals are rectified and integrated during the gated period of the rising edge of each burst. The resulting integrated values then correctly represent the relative displacements of successive different parts of the tire wall to be inspected. In one exemplary embodiment, a common average is calculated from successive observations at each ring wall position to avoid possible standing wave zeros and the like that may occur at a particular frequency and ring geometry. Such values can be displayed on the CRT for inspection and fault detection. Alternatively, such values may be digitized (possibly by a nonlinear exponential A-D process to improve the effective signal-to-noise ratio at relatively low signal strengths) prior to display and / or processed by I) 7 7281 6 recognition algorithms of the desired format on a digital computer to automatically separate ring anomalies, such as layers.

It has also been found that improved performance results when acoustic signals have a relatively high frequency (e.g., greater than approximately 40 kHz and in a preferred embodiment 75 kHz). Such moderately high acoustic frequencies tend to avoid unwanted erroneous indications caused by common ambient acoustic sources and at the same time have relatively small wavelengths (e.g. approximately 3.8 cm in tire rubber), which improves the resolution of relatively small tire defects without. unnecessary complication in the observed shift numbers due to such a small wavelength that the signals could be affected by anomalies in the ring structure that are not actual errors.

Averaging the received signal over several periods during the leading edge of each burst improves the signal-to-noise ratio of the resulting measured values, as does the use of a nonlinear A-D process. Meaning the values obtained at different frequencies can further improve the results.

The use of an inflated tire in a preferred embodiment has been found to help maintain the actual surface of the rotating tire and thus avoid variations that could otherwise be caused by "throwing" the tire or some other axial movements of the tire walls during rotation. An air-filled tire is also useful in that it helps to at least partially stress the walls of the tire as they are during normal use and to open leaks through the walls of the tire so that they can be detected by ultrasonic detection as air passes through them. A pressure of approximately 35 kPa is required to maintain a stable air-filled tire structure. However, it has been found that improved signal transmission and improved overall performance are obtained if the ring is inflated with air in the weight range of approximately 100-125 kPa.

7281 6 8

It is preferred, although not necessary, that the tread of the tire being inspected be first ground to obtain a uniform surface, which reduces erroneous defects that might otherwise be caused by tread patterning and / or uneven wear points or patterns on the tire's tread.

In this connection, it should be mentioned that the tire grinding equipment and the present method can be used as a combined, comfortable and efficient overall operation. In any case, since such a grinding operation is inevitably included in the tire retreading operation, this combination is particularly attractive when the tire carcasses are inspected in preparation for retreading.

A preferred exemplary embodiment of the present invention also includes specific mechanical properties for automatically moving acoustic transducers to and from the operating position relative to the walls of the inflated ring. During tire attachment and detachment, the acoustic transmitters are retracted both radially and axially relative to the at least one tire attachment ring or flange for the purpose of both facilitating tire installation and detachment operations and protecting the acoustic transmitters from possible physical damage. During or after inflating the tire, these acoustic transmitters are moved radially outwardly inside the inflated tire to an operating position relative to the inner surface of the tire walls. At the same time, the system of acoustic receivers is moved radially inwards towards the outer surfaces of the walls of the inflated ring to the desired operating position.

In a preferred exemplary embodiment, the relative axial movement of the acoustic transmitters relative to the tire mounting flange or ring is accomplished by spring loading the tire retaining ring so as to move axially away from the acoustic transmission, thereby exposing them to the ring Such a spring load also helps the tire rims to properly position the mounting wires or rings during the tire mounting and inflation operations.

Il 7281 6 The ultrasonic bursts and the reception gate cycles are preferably synchronized to hit corresponding successive incremental points on the rotating ring so that the final display or error indication can be accurately located relative to a mark on the ring and / or tire mounting flange or the like.

The foregoing and other objects and advantages of the present invention will become more apparent from the following detailed description taken in conjunction with the accompanying drawings of a preferred exemplary embodiment thereof.

Figures 1 and 2 are perspective views of a combined NDI / grinder constructed in accordance with the present invention.

Figure 3 is a block diagram of the invention shown in Figures 1 and 2. Figure 4 is a block diagram of ultrasonic NDI circuits that may be used in the NDI / grinder of Figures 1-3 or in a machine having only NDI capability.

Fig. 5 is a schematic diagram illustrating a ring wall portion, acoustic transmitters and receivers, and the preamplification and multiplexing circuit shown in Fig. 4.

Figure 6 is a detailed circuit diagram of the preamplifier shown in Figure 5.

Figure 7 is a detailed circuit diagram of a representative individual of the signal processing channels shown in Figure 4. Figures 8a and 8b are detailed circuit diagrams of the system interface shown in Figure 4.

Fig. 9 is a detailed circuit diagram of the central processing unit (CPU) shown in Fig. 4.

Fig. 10 is a detailed circuit diagram of the display interface shown in Fig. 4.

Figure 11 is a schematic representation of several representative waveforms useful in explaining the operation of the circuits illustrated in Figures 4-10.

Fig. 12 is a cross-sectional view of a keel climator / impedance matching device used in each receiving transducer.

Figures 13 and 14 show curves taken from the CRT outputs obtained in a non-destructive inspection of a ground wheel body in accordance with the present invention.

7281 6 10

Fig. 15 is a detailed sectional view of the fixed mandrel and transmitter mounting arrangement of the embodiment of Figs. 1 and 2.

Figures 16 and 17 are flow charts of suitable control programs for use in the CPU of Figures 4-10.

Fig. 18 shows a circuit diagram of the second circuit for providing the AVS amplifier and integration channels, and Figs. 19, 20a and 20b show a program flow diagram for first searching for air leaks and then disconnections in two groups of eight channels.

Figures 1 and 2 are thus two perspective views of a preferred exemplary embodiment in which a grinding and NDI machine is combined. As is obvious, the NDI properties of the machine can be used without grinding capability if desired.

The larger mechanical components of the machine are mounted on an open frame 100 with a fixed mandrel 102 and an axially movable mandrel 104 facing each other on the same horizontal shaft 106. Conventional circular tire mounting rings or flanges 108 and 110 are attached to the rotatable outer ends of the mandrels 102 and 104. 112 between them. The conventional pneumatically operated tire lifting mechanism 114 is preferably used to assist the user in lifting and turning the tire into and out of the mounting rings 108 and 110 during and after mounting and removing the tire.

The mounting ring 108 and thus the wheel ring 112 is driven by a two-horsepower DC motor 116 through a reduction gear 118. A tire surface speed of approximately 183 m per minute is preferred for grinding operations, while a much lower speed of approximately 12 m per minute is preferred for NDI operations. The spindle 104 and thus the mounting ring 110 is pushed axially forward and retracted by the pneumatic cylinder 120. During the mounting operation of the tire, the mounting ring 110 is retracted by the cylinder 120 to allow the ring 112 to be lifted into place on the mounting ring 108. The mounting ring 110 is then pushed against the respective circumference ring 112 is inflated with air

K

11 7281 6 to the pressure through the center of the spindle 102.

A conventional rotating tire grinding grid 200 is mounted on a stand 202 located at the rear of the machine, as seen in Figure 2. The rasp 200 is guided by a conventional panel 204 to move transversely along the desired grinding path 206 and horizontally toward and away from the ring. Control is provided by a conventional control mechanism, including a joystick used to control the pneumatic cylinder 208, guide screws, and associated drive motors, and the like. The grinding grid 200 is rotated by a separate motor mounted on a stand 202. The grinding mechanism itself is of a conventional type marketed, e.g., by Bandag Inc., e.g., Buffer Model No. 23A.

The system (row) of 16 acoustic ultrasonic receiving crown detectors 210 is located above and around the outer walls of the ring 112. The receivers 210 preferably include a conically formed collimator and / or focusing tube to help limit the receiving field of each individual transducer to a relatively small and unique area across the ring wall. The receivers 210 may conveniently be individually or in groups canned in polyurethane foam or the like to assist in mechanically securing the receivers to their desired positions, to help protect the receivers, and to help isolate the receivers from interfering external acoustic signals. The row of receivers 210 is radially set to the operating position by the air cylinder 212 with the hydraulic steering cylinder connected, for the purpose of determining a radially advanced operating position for the receivers 210.

A block diagram for a combined tire grinding / NDI machine and associated electrical and pneumatic circuits is shown in Figure 3. The electric motor and pneumatic cylinder controllers 300 are quite conventional and are therefore not described in detail. The user inputs, indicated on the left in Figure 3, are made directly or indirectly by the user through conventional electrical switches, relays, air valves and / or fluid control valves.

7281 6 12

When the machine is in operation, the wheel ring is placed in the lifter 114 and lifted into position between the mounting rings 108 and 110. Preferably, the predetermined pointer point of the ring is aligned with the pointer point on the flange 108. The fastening apparatus is then engaged by causing the flange 110 to move into the ring 112 so as to pinch the ring flanges together in preparation for inflating the tire. The tire is then inflated to the desired set pressure value. As will be explained in more detail below, the flange 108 is spring loaded so that during attachment and inflation of the ring it is caused to move axially outward against the spring load (e.g., approximately 5 cm). This facilitates the ring filling process and at the same time exposes the ultrasonic transmitter located inside the ring from a relatively filtered position so that it can then be pushed out to the operating position below the row of receivers 210. A locking switch activated by air pressure and / or physical movement of the flange 108 can be used to prevent the transmitter from protruding prematurely before it is exposed from its protected position.

In the grinding mode, there is no need to push the transmitter out. The drive motors of the grinding tarasp are activated and controlled in a conventional manner (e.g. with a joystick and conventional pushbutton controls) to grind the tread of the tire as desired. It is preferred, although not necessary, to grind the tire to a substantially flat tread before the NDI operation is performed. Such grinding is believed to prevent possible erroneous indication indications that could be caused by normal tread patterns and / or uneven tire surface wear.

When the user selects the NDI mode, the ultrasonic transmitter inside the inflated ring 112 is pushed into the operating position and the row of receivers 210 is lowered into the operating position by the pneumatic cylinders associated with the row. The speed of the same 2-horsepower DC motor, which operates the tire at a surface speed of approximately 180 m per minute during grinding operations, can be reduced with conventional electrical circuits to operate the tire at a speed of approximately 12 surface meters per minute during the NDI mode. After

II

7281 6 13 movement has reached a serious state, the user can activate the radar input switch for the ultrasonic NDI circuits 302. The tire walls are then ultrasonically inspected during one full turn of the wheel to produce a display 304 that can be directly interpreted or indirectly indicated by a person (e.g. still satisfactory for grinding and coating, suspicious magic unsatisfactory). If the condition in question is indicated, the tire may be discarded or further refined and then retested.

The ultrasonic NDI circuits 203 are shown in more detail in Figures 4-10. As seen in Figure 4, the outputs 16 of the ultrasonic receiver 210 are amplified and multiplexed into eight signal processing channels AH by circuits 402, shown in more detail in Figure 5. Each signal processing channel then provides AVS (automatic gain control), rectification, integration, and integration. A representative channel of this processing field is shown in detail in Figure 7. The resulting digitized outputs are shown on a conventional eight-bit data bus 406 connected to a conventional microprocessor CPU (e.g., an 8080 type 8-bit processor). and connected via data bus 406 to data memory 412, programmable read only memory (PROM) 414, and system interface circuit 416, shown in detail in Figure 8. Display interface 418 (shown in detail in Figure 10) is su. connected to data memory banks 412 to produce a CRT-type oscilloscope.

The system interface 416 provides the necessary gating and other control signals to the signal processing circuit 404 and also provides HIGH CHAN (high channel) multiplexing signals to the preamplifier circuits 402 and the transmitter drive circuits and the multiplexing circuit 422 used to operate the ultrasound. The operation of the entire system is synchronized with the rotational movements of the ring 112 by a rotating pulse generator 424 which is driven directly by the ring (e.g., geared to the reduction stage). The rotary pulse generator 424 produces 1024 pulses per revolution in pin RPGX and 1 pulse revolution in pin RPGY.

7281 6 14

As shown in Figure 5, the acoustic ultrasonic transmission crystals 500 and 502 are housed within an air-filled ring 112 mounted between flanges 108 and 110, which in turn are rotatably mounted on spindles 102 and 104. The electrical conductors supplying transmitters 500 and 502 are routed to a fixed spindle 102 to the transmitter activation circuits. The filling air is likewise passed through the center of the mandrel 102 as well as the pneumatic lines and / or other control connections for pushing out and retracting the transmitters.

The exemplary ultrasonic transmitters 500 and 502 have a radiation field substantially comprising a sector of approximately 90 °. Therefore, they are mounted at an angle of 90 ° to each other in a block 504 which may be formed of, for example, polyvinyl chloride plastic. It has been found that acceptable operation does not result if the transmitters are too close to or too far from the inner surfaces of the ring. In a preferred exemplary embodiment, the transmission crystals 500 and 502 are approximately 5 cm from the inner surfaces of the ring, although this optimal distance can be varied considerably (e.g., plus or minus approximately 2.5 crar).

The in-line receiving transducers 210 are arranged in an arc that generally corresponds to the shape of the outer side of the ring wall. Here, too, it has been found that acceptable operation does not result if the receivers are too close to or too far from the outer walls of the ring. Preferably, the receivers are no closer than approximately 2.5 cm from the outer surface of the ring, but are preferably 14-22 cm from the opposite transmission crystal. The receiving transducers 210 each preferably use a conically shaped collimator and / or focusing tube, as detailed in Figure 12. These tubes are preferably machined from polyvinyl chloride plastic and also help to match the impedance of the transducer crystal surface used to the acoustic impedance of the ambient air.

A moderately high ultrasonic frequency is used to help avoid interference from erroneous surrounding acoustic signals and to achieve improved resolution by using shorter wavelength acoustic signals while avoiding ultra high frequency acoustic signals and associated problems. Frequencies above 40 kHz are recommended, 75 kHz is currently the preferred preferred frequency. Ultrasonic transducer crystals operating at 75 kHz are commercially available. Examples of available receiving crystals include the MK-111 transducer from Massa Corporation, Windom, Massachusetts, which has the following values:

Maximum impedance frequency (fml 75 kHz + 3 kHz

Impedance at fm (min) 6 points

Receiving sensitivity (O.C.) at maximum output frequency dB in relation to 1 volt / microbar -70 dB min Transmission sensitivity in relation to 1 microbar at 30 cm / 10 mW -12 dB min

Maximum power input 100 mW

Directivity -10 dB at a maximum total angle of 90 °

Thermal stability 10 Ti change in frequency

in the range of -34 ° C to + 66 ° C

Capacitance 120 + 20% pF

A suitable transmitting crystal tuned to approximately 75 kHz is available from Ametek / Straza, California, 8-6A016853.

The electrical conductors from each transducer 210 are preferably connected in coaxial cables 506 to associated preamplifiers 508. The outputs of each of the 16 amplifiers 503 are connected to an eight-pin two-position electronic switch formed by a Signetics SD5000 integrated circuit controlled by a HIGH CHAN multiplexing signal 41 the eight mug-plated output channels are connected to the process channels AH via transistor buffer amplifiers. Accordingly, in the absence of the HIGH CHAN multiplex signal, the outputs of the first eight preamplifiers 5C8 are connected to the corresponding sic? Signal processing channels A-H. However, when the HIGH CHAN multi-16 7281 6 plexing signal is present, the outputs of the last eight preamplifiers 508 are connected to the corresponding signal processing channels A-H.

The connection of each preamplifier 508 is shown in more detail in Figure 6. It includes a first transistorized stage with a gain of approximately 150, followed by a cascaded integrated circuit amplifier with a gain of approximately 11.

The signal processing circuits 404 for each channel A-H are identical. Accordingly, the circuit for channel A is shown in Figure 7. The waveforms shown in Figure 11 are useful for understanding the operation of the circuit of Figure 7.

Generating a pulsed or burst ultrasonic waveform to operate the transmitters 500 and 502 is explained below. However, it can be seen from Figure 11 that each transmitter is used to produce at least one burst of approximately 50 cycles of 75 kHz acoustic output signals each time an RPGX trigger pulse occurs (e.g., 1024 times per wheel rotation). After the transmission delay, which depends on the distance between the transmitter and the receiver and between the characteristics of the outdoor air and the rubber of the tire, the transmitted acoustic signals are received. The received and converted acoustic signals can have a complex amplitude diversity curve (and not as well-behaved as shown in Figure 11) depending on the type of multiple reflections, internal reverberations, wave cancellations, and / or other strange wave effects that occur along the paths. As a result, only the leading edge or the initial part of each such uitronic pulse or burst (e.g., where the amplitude envelope initially grows) provides the best and most accurate indication of the quality of the transmission path (i.e., the ring errors involved). As a result, the signal processing circuit shown in Fig. 7 is adapted to efficiently utilize only such an initial or leading edge portion of each burst of ultrasonic signals. In one embodiment, the data for each ring measurement area is obtained by averaging measurements taken at different frequencies of the acoustic signal.

li 7281 6 17

As disclosed in U.S. Patent 3,882,717, it is necessary to automatically provide a gain-adjusted gain from the transmission of ultrasonic test signals to compensate for different average thicknesses of the tire carcass. In this previously patented system, only one signal processing channel with AVS was used to compensate for differences in the average thickness of the ring carcass of a given tire cross section. However, it has been found that an automatically gain-controlled gain must also be included in each of the plurality of test channels of this invention to compensate for differences in the average thickness of the tire carcass from ring to tire.

Accordingly, the AVS amplifier 700 (e.g., the integrated circuit MC1352) was not included in the channel A shown in Fig. 7.

The ultrasound signals passing through channel A are fed back to pin 10 in AVS amplifier 700 and the input to RC circuit 702 having a relatively long time constant (e.g., 10 seconds) is connected to pin 9 in amplifier 700. As a result, the averages of the signals passing through channel last over several seconds (including periods when the amplifier is capable) is compared to a reference AVS constant bias voltage at pin 6, with the intention of maintaining a substantially constant average output level at pin 7 for an RC time constant period. In a preferred exemplary embodiment, amplifier 700 has a gain that can automatically range from 1-1000.

Amplifiers 704 and 706 are cascaded in channel A and each provides a gain of approximately 2. In addition, diodes 708 and 710 are coupled to amplifier 706 to cause its output signals to be rectified as shown in FET gate 712.

Returning to Figure 11, the integrated reset signal INTGRST is generated during a first transmission delay period at a particular test ring position and allowed at FET port 714 (Figure 7) to discharge an integration capacitor 716 connected across amplifier 718 (forming a Miller-type integrator). In addition, the AVS amplifier 18 7281 6 700 becomes capable of an AGCEN signal at some point during each test period to take a sample of the received signals. The integrator-enabling signal INTGEN is timed to enable the FET switch 712 only during the beginning, i.e., the leading edge, of the ultrasonic burst (e.g., for approximately 130 microseconds, i.e., the 75 kHz burst for approximately the first ten cycles). If desired, two or more received bursts at different frequencies can be aggregated and the results integrated together to efficiently average measurements at different frequencies (and therefore at frequencies with different standing waveforms).

The output of the integrator 718 is then converted to a digital signal under the program control of the CPU 408, generating suitable analog DAC inputs to the comparator 720 and conversion-port signals to the CONV port 722, which interfaces with one of the conventional data bus lines (in this case DB0). Such program-controlled analog-to-digital conversion is conventional and involves the conversion of a CPU program-controlled reference digital signal to an analog DAC signal, which is then sequentially compared in comparator 720 to the results of such comparison made available to CPU via data bus lines and ports 722. For the various signals known by the sequential comparison process, the programmed CPU is able to determine a digital value corresponding to the integrated analog value of the input from the amplifier 718.

This process is, of course, repeated simultaneously in channels A-H and sequentially in each channel for each burst or group of bursts of ultrasonic signals occurring at a particular wheel wall test site.

Figure 8 shows that the RPGX (1024 pulses per revolution) and RPGY (1 pulse per revolution) signals from the rotating pulse generator are routed through the three-state buffers 800 to the data bus lines DB0 and DB1 depending on the address signals generated by the CPU3 and Q4. Other address output inputs from the CPU are input to output decoder 802 to produce signals ULOS320.00 to ULOS320.70 under proper program control.

7281 6 19

Just before the probe cycle, the CPU is repeatedly programmed to examine the DB2 data bus line for the probe query signal SCANRQ generated by the user by manipulating the probe query switch 804, causing the flip-flop 806 to be set ULOS320,60 the next time it occurs.

After the scan request is detected by the CPU via the data bus line DB2, the CPU is programmed to examine the RPGX and RPGY signals, which are then applied to the data bus lines DB0 and DB1 by the address inputs IN3 and Q4. The actual measurement period will not start until a second RPGY signal is detected. The purpose of this is to ensure that the ring rotates correctly at a substantially constant speed and that the AVS circuits function properly. Thereafter, with each detected RPGX signal, the CPU is programmed to generate a signal ULOS320,10. The ULOS320,10 signal triggers the monostable circuits 808 and 810 and also makes the latch 812 capable of accepting the digital values represented by the data bus lines DB0-DB4.

Just before generating the first burst of ultrasonic waves at a particular test point on the ring wall, the CPU generates ULOS320,70, which triggers a reset to monostable member 322 and produces an integrator reset signal via flip-flop 823 and an AND gate 825.

The 4-bit binary counters 814 and 816 are cascaded to count the 18.432 MHz clock signal inputs from the CPU and divide these clock pulses by a numerical factor represented by the contents of the cache 812. The result is an approximately 75 kHz clock signal (both 74 kHz and 76 kHz frequencies are used sequentially in one embodiment and the two results are averaged), which is used to trigger a monostable member 818 having an adjustable time period so that its output can be set substantially a square wave. 50% as a voltage cycle signal. As shown in Fig. 8, the monostable member 818 is controlled by a pulse amplification signal from the addressable flip-flop 319. If desired (e.g., to listen for leaks), the ultrasound outputs can thus be selectively disabled by the CPU.

20 7 2 8 1 6

An approximately 75 kHz 50% voltage period signal is then buffered through amplifier 820 and represented as a square wave-out input MB (see Figure 11) to conventional transmitter drive amplifiers (producing approximately 200 volts from peak to peak electrical output), which in turn cause a generally sinusoidal 75 kHz acoustic output from the transmitter, as shown in Figure 11.

This MB generation of approximately 75 kHz output continues until one shot of the 808 is timed out (e.g., approximately 1 millisecond). During this period, a burst of ultrasonic acoustic signals is transmitted from one of the transmission crystals.

The time of the monostable body (one shot) 810 is set to delay approximately as much but slightly less than the transfer delay between acoustic transducers. The delayed output from member 810 resets the data-ready flip-flop 828 and Touches the integration timing to the monostable member 826, which produces a signal INTGEN that enables integration. At the end of the integrating signal from the member 826, the data-ready flip-flop 828 is set to produce a slip-ready signal via the data bus line DB4 to the CPU.

If more than one analog data value is to be combined at the integrator output, the CPU is simply programmed to ignore the data-ready signal until the required number of measurement cycles has been completed. Ultimately, however, the data-ready signal indicates to the CPU that the analog-to-digital conversion of the integrated analog signal is now ready to be performed. The CPU, under conventional program control, then begins to produce various analog reference signals from the DAC digital-to-analog converter 830, controlled by digital data cached on the data bus lines 832 with the address signal OUT320.00. At the same time, the CPU is programmed to produce the correct conversion-gating signals via the CONV address inputs to ports 834, 836 and 833.

The DAC may be a non-linear type 08 or a non-linear exponential type 76 or a non-linear DAC coupling of some other known type. Nonlinear DAC-76 is believed to improve the effective segregation-interference ratio at lower

II

2i 7281 for 6 levels of signals.

The CPU is normally programmed to produce a multiplexing HIGH CHAN output by setting and resetting the addressable flip-flop 840 via address lines A0 to A2, OUT320.3o according to the data value then present in data line DB0. However, the manual bypass switch 842 is positioned in the circuit so that either low channels 0-7 or high channels 8-15 can be manually forced through three-state buffers 844, the outputs of which are connected to data bus lines DB6 and DB7.

A flow chart of an example CPU control program is shown in Figures 16 and 17. Conventional power switching, reset, and start-up steps are described in block 1501. After the START input (output), the flip-flop 806 (Figure 8) is reset, integrators are incapable, via flip-flop 323 Fig. 8) and the data memory circuits are incapable in block 1503. A round-trip loop 1505 is then entered and held until SCANRQ is detected in DB2.

After the radar requirement is detected, the indicator lamps are tested, the integrators are enabled for normal operation (via flip-flop 823), the data memory is enabled for CPU access (and vice versa, the display interface is disabled for normal memory access) in block 150 7. High 'high' switch 842 (Figure 8) is also checked via DB6 and DB7. If a low or normal mode is indicated, the HIGH CHAN multiplex signal is considered zero through flip-flop 840. It then goes to kvselysil to test the transfer of 1505 RPGY. A similar interrogation loop 1510 is then entered to output at least one ring rotation before the measurement results are taken. The programming counter Θ ^ is then set to zero and loop 1 enters the test subroutine (Figure 16). As will be explained in more detail below, step 1 is performed 1024 times to collect and store 1024 data values in each of the eight transducer patterns corresponding to 1024 tire test points divided by the entire 360 ° rotation of the ring in each of the eight channels.

7281 6

After entering loop 1, the RPGX signal on line DB0 is tested for a 1 to zero shift in loop 1600. After this transition occurs, all integrators are reset via monostable element 822 (Figure 8), cache 812 is set to produce a 74 kHz MB operating signal, and transducers are used a burst of 74 kHz MB of the operating signals through the monostable element 808 and a signal enabling the pulse generator via the flip-flop 819. Because the monostile member 810 is also Touched, the leading edge of the received burst becomes gated and integrated in each channel.

While this test is being performed at 74 kHz, the CPU is in the wait loop 1602. The cache 812 is then returned to produce a 76 kHz MB signal and the transmitters are given pulses again. The result is a second gate-integrated integration at the leading edge of the received burst at 76 kHz. As soon as this second integration is completed, a data-ready signal is detected in DB4 in the wait loop 1604.

After the analog information has thus been collected at two different frequencies at a particular test point in the ring, the AVS circuits are keyed (to keep them active by sampling the channel signal level for the relevant RC time constant) and a conventional analog-to-digital conversion is introduced. This routine converts each integrator output to a six-bit digital value, which is then stored in data memory 412. Data for each channel is stored in a different portion of memory so that similar data points for each channel can later be assigned using the same lower class memory address signals.

The current of the software glass is then made to take one step and return to loop 1, unless data measurements have already been taken from all 1024 tire test points.

After the first exit, loop 1 may, if desired, go to the pattern recognition subroutine in block 1513. The pattern recognition results may then be tested at 1515 and 1517 to determine which of the status indicators 846 (Figure 8) is to be turned on. Alternatively, it can be

II

7281 6 23 pattern recognition steps are skipped, as shown in the figure by dashed line 1518, to trigger a HIGH CHAN multiplex signal if the operation is in normal mode. (If only high or low channel testing is forced with switch 842, a return to the START input point can be made). Measurements are then made for a higher group of eight channels, as is apparent.

Although loop 1 in Figure 17 causes the 74 kHz and 76 kHz measurements to be combined, it is also apparent that block 1606 can be bypassed if measurements at only one frequency are desired. In the same way, measurements on more than two frequencies can be combined if desired. In addition, the combination of several data values can initially be done either in analog form (such as in the exemplary embodiment) or in digital form, which is obvious in the light of the above.

As already mentioned, the CPU may, if desired, be programmed to automatically analyze the digitized data collected during one complete sampling period by pattern recognition algorithms and activate one of indicator indicators 846 (e.g., accepting, rejecting, or leaking) through conventional lamp drive circuits 848 controlled by cache 850 contents. which are filled on the data bus lines DB0 to DB4 under the control of an address-integrated ULOS320.20 signal. Air leakage can be detected, for example, by performing a complete sampling and measurement cycle while disabling the ultrasonic transmitter gaps. The observed increases in the received signals are then detected as leaks.

The central processing unit shown in Fig. 9 is conventionally connected to decode the various address lines and to produce the already mentioned address inputs to the system interface shown in Fig. 8. The CPU itself is a conventional microprocessor formed by an integrated circuit 8080 having data input and data output lines D0 to D7 connected to data bus lines DB0 to DB7 by conventional bidirectional common and station circuits 900. The address lines A0 to A9 and A13 directly connected to system interface, memory circuits, etc. Address cables AIO, Ali and 7281 6 24

Al2 is decoded in decoder 904 to produce address generation outputs Q0 to Q7. Similarly, the address lines A14 and Ai5 are decoded together with the normal write and data bus input signals from the CPU in the decoder circuit 906, which outputs IN0 to IN3 and OUT0 to OUT3 address output. The normal input CPU signal DBSIN and the address generation line 814 and 815 are also connected via gates 908 and 910 conventionally to provide directional power input to bidirectional common line stations 900. Around 18 MHz clock 912 is also conventionally connected to the 8080 CPU. However, pin 12 of the integrated circuit 3G8224 has been output to provide a clock frequency of 18.432 MHz to the frequency dividers in the system interface already discussed in connection with Figure 8.

The data memory circuits are conventionally formed by combining 25 integrated circuits of the 4045 type to produce 8,192 eight bit word (or byte) data memory capacities.

Programmable read only memories (read-only memories) may be formed of three integrated circuits of type 2708, each of which constitutes a programmed memory of 1024 bytes. The read / write memory of the 256-bit word is also preferably connected to the CPU as part of programmable memory circuits. A type 2111-1 integrated circuit can be used for this purpose.

The CRT display interface is directly connected to the data memory table.

After the entire measurement period is completed (e.g., when the third RPGY signal is detected after the scan request), there are 1024 data values to use for each of the 16 measurement channels, representing the relative magnitudes of the ultrasonic signals transmitted through the ring in 1024 consecutive at a corresponding point in the circumferential region of the ring monitored by the receiving transducer for a particular channel. This digital information can be converted to conventional video usage signals for the CRT and displayed as shown in Figures 13 and 14. Alternatively, the 8080 processor may be programmed to analyze (e.g., Figure 1).

II

7281 6 25 identification algorithms) available digital information and to activate the appropriate indicators for Hemp 846, shown in Figure 8.

The display interface shown in Figure 10 is conventionally connected directly to the data memory 412 via the memory data buses 1000, the memory quarter selection via the common lines 1002, the nuisance address common lines 1004, and the data cache via the pulse line 1006. The entire display may optionally be enabled or disabled, as desired, under the control of the CPU via the CPU addressing outputs A1, Q3, OUT3 and A0, and through the flip-flop 1008 and associated inverter and ports shown in Figure 10. In a preferred embodiment, the display interface is disabled, at least when other parts of the system have access to the data memory 412 for the purpose of preventing possible simultaneous activation of the data memory circuits.

The display interface is operated at 11.445 MHz by a clock 1010. Its output uses a counter 1012 connected to divide the clock signals by a factor of 70. The first 64 counts of the counter 1012 are used by a comparator 1014 which also receives 6 bits of data (i.e. 64 different numeric values) from the assigned data memory location. representing the magnitude of the ultrasonic signals transmitted through a particular test point in the ring.

Thus, the output of comparator 1014 on line 1016 occurs at a predetermined time within 64 clock cycles corresponding to the magnitude of the input-input digital information over lines 1000. The clock pulse during data concurrency causes a momentary shift of flip-flop 1013 and produces a video output pulse through port 1020, which pulse has a time width and spacing of one display point corresponding to the amount of information stored within the corresponding channel time slot. The flip-flop 1022 becomes triggered by the counter 1012 when counting the 65th clock pulse and generates an in-channel discriminating shutdown video pulse from the port 1020. The counter 1012 then continues counting down another 5 clock pulses before resetting itself and starting a new cycle using information from the next channel.

The 70th count from the counter 1012 also uses a three-bit channel counter 2624 which, through three to eight decoders 1026, sequentially addresses eight different portions of the data memory corresponding to eight of the sixteen ultrasonic channels. The choice between higher or lower eight channel displays is made with switch 1028.

At the end of the complete horizontal scan line, 10 x 70 clock pulses (2 x 70 clock pulses are counted during horizontal reset) are counted by counters 1012 and 1024 and the transfer pulse goes to 12-bit counter 1029 to step the addresses on line 1004 (via decoder 1030) for the next horizontal offset line. In the case of a standard interleaved CRT scan raster, virtually every other horizontal line is skipped and these are highlighted in the second horizontal raster in a known manner. The states of the counters 1024 and 1029 provide all the necessary timing information for generating standard CRT horizontal synchronization, vertical synchronization, and vertical and horizontal reset shutdown video signals in the member 1032.

The various video signals are conventionally mixed in a video amplifier 1034, the output of which goes to a CRT display.

Because each channel has 1024 data values, but the standard CRT raster has much fewer horizontal lines, switch 1036 is used to select only odd or even addresses for data values on a particular channel. Thus, the full 360 ° examined tire surface, in a given channel, is displayed in the relevant time slot by 512 horizontal scan lines spaced vertically apart.

As explained, the data values of a particular channel may be divided into a vertical portion of the CRT display and shifted horizontally from the vertical basic data line according to the stored data values. However, in the preferred embodiment, the CRT deflection coil is rotated 90 ° so that the final CRT display of the channel is shown horizontally, as shown in Figures 13 and 14.

As shown in Figures 13 and 14, the signal lines in each individual channel deviate upward to describe the reduced

II

7281 6 27 ultrasonic signal quantities. Accordingly, it can be seen from Figure 13 that errors have occurred in channels 12 and 13 approximately 20 ° from the indicator mark. Similarly, an error is indicated in Figure 14 in channels 12, 13 and 14 at approximately 280 °.

Although not shown in Figures 13 and 14, if leakage occurs, this would be indicated as an increased signal magnitude, which in the representation of Figures 13 and 14 would have resulted in a downward deviation of the signal line in the corresponding channel.

The lines in channels 0 to 3 and 12 to 15 are caused by the ends of the wire, the transition between different normal tire layers, and the periodic patterning of the remaining tread structures at the outer edge of the tire tread. The information actually shown in Figures 13 and 14 was taken using a linear DAC circuit in the analog-to-digital conversion process.

In more detail, the fixed mandrel 102 and the associated transmitter mounting arrangement are shown in section in Figure 15. The transmission crystals 500 and 502 are oriented at a 90 ° angle to each other from the surface of the PVC mounting block 1500. The block 1500 is in turn attached to a retractable rod 1502 connected to the piston of the pneumatic cylinder 1504.

As shown in Fig. 15, the pneumatic cylinder 1504 has pulled the transmission crystals 500 and 502 back to a protected area bounded by an annular plate 1506 attached to the tire mounting flange 108. The tire mounting flange 108 is rotatably mounted to the fixed mandrel 102 by ball bearing compositions 1508 and 15. the coupling is kept airtight by a rotating sealing assembly 1512. The center of the mandrel 102 is hollow to pass through the electrical conductors of the pneumatic control line 1514 and the transmitter.

The rotating flange 108 and the composition associated therewith are loaded by a spring 1517 toward its axially extended position, as shown in Figure 15. However, the flange 108 can be axially displaced against the spring force to a position 7281 β 28 not shown in broken lines in Figure. In a preferred exemplary embodiment, such movement begins when the flange 108 is subjected to a horizontal force of approximately 680 kg (14 kPa). A sliding joint that allows this movement is also kept airtight by the O-ring 1516. In the exemplary embodiment, no axial movement greater than approximately 5 cm is allowed until the spring force is sufficient to resist further movement even when the ring is inflated to approximately 100-125 kPa atmospheric pressure.

When the flange 108 is moved to the left, as shown in phantom in Figure 15, against the force of the spring 1519, the exposed transmitters 500 and 512 and the pneumatic cylinder 1504 can be activated to push the transmitter unit to the position shown in phantom in Figure 15 for the functional measurement period. Suitable locking switches activated by the pressure inside the inflated ring and / or the physical axial position of the flange 108 can be used to ensure that the pneumatic cylinder is not inadvertently activated to push the transmitters outward and damaged while the transmitters 500 and 502 are still enclosed and protected by the flange 1506.

Figure 18 shows a second circuit for generating an AVS amplifier and integrator channels. This coupling allows the generation of INTGEN, AGCEN, INTGRIST, and MBT pulses from RPG pulses at 1605 or the generation of simulated RPG pulses from an addressable cache 1608 under program control.

When the RPG simulator is capable, there is a 1608 output labeled 5, 50% duty cycle pulse sequence selected by the multiplexer 1611 To tap the monostable members 1612 and 1613. The monostable member 1612 is triggered at the rising edge of the 1611 output and times to 300 ns . The monostable member 1613 becomes triggered at the falling edge of the I611 and also timed out at 300 ns.

The outputs of the elements 1612 and 1613 are connected to a touch mono-stable delay element 1614 and a monostable element 1615. 75 kHz

II

The development of bursts by members 1615, 1620, 1621, 1622 and 1623 has been previously described above. Monostable Delay Member 1614 Touches the monostable integrator 1616 and returns the data-ready flip-flop 1617.

Flip-flop 1617 signals that the AVS amplifier / integrator channel outputs are ready for digitization. Flip-flop 1617 will only be set when the RPG is high.

Flip-flop 1617 Touches the AGC flip-flop 1619, the level of which is shifted and sent to the AVS amplifiers.

The delayed RPG signal occurs at the output of flip-flop 1618 and is used by the software to synchronize with the wheel rotation.

When the simulator is disabled, the multiplexer 1611 sends the logic output of the member 1605 to the monostable members 1612 and 1613. The input source for the member 1611 now comes from the RPGX pulses generated by wheel rotation and the necessary outputs, i. INTGEN development is accomplished by directing the output pulses of multiplexer 1611 to monostable members 1612 and 1613.

The sequence of triggers of monostable organs follows the same shape as explained in the previous paragraphs when the RPG simulator is activated.

The DAC formed by the elements 1624 and 1625 generates an analog voltage which is used by the CPU for analog-to-digital conversion of the integrated values of the received signals.

Decoder 1609, flip-flop 1610, register 1627, and lamp drive 1628 perform the functions already described above. The cache memory decoder 1629 and display element 1630 provide status information during program execution.

During air leak detection, the PULSE generated by the software member 1608 is low, thus preventing MB activation pulses in the pulse unit by emptying the monostable member 1620.

. 30 7281 6

Figures 19, 20a and 20b illustrate a program sequence that searches for air leaks and then searches for detachments in two groups of eight channels.

Blocks 1631 and 1632 are initial states for the system and block 1633 selects an RPG simulator to tap the monostable timing elements. The RPG simulator is capable of alternating high and low speeds of 8 ms while the SCAN RQ flip-flop is tested in loops 1634 and 1635. The RPG simulator upgrades the AVS levels so that when SCAN RQ becomes active, data acquisition on air leaks can begin immediately.

When the SCAN RQ becomes active, the RPG unit is selected in block 1636 and the data memory is enabled in block 1637. The subroutine to retrieve data is called in block 1638 and this is detailed in Figures 20a and 20b. Next, the pattern detection is called in block 1639 and air leaks are detected and the air leakage lamp is connected in blocks 1640 and 1641.

The pulse cell is now activated in block 1642. Tests for high channel, low channel only, and normal probe are performed in blocks 1643, 1644, and 1645.

The COMPLETE INFORMATION and PATTERN IDENTIFICATION subroutines are called in block 1646. Blocks 1647, 1648, 1649, and 1650 test the reject / accept state and decide whether to continue scanning in the high channel group. Fetch DATA and PATTERN IDENTIFICATION is called again in block 1651 and the tire status is tested again in block 1652 and the program returns continues through block 1653, "reject" status, or block 1654, "accept" status.

Figures 20a and 20b show a detailed flow chart for the Fetch INFORMATION subroutine. Position counter, Θ current is set to zero in block 1655. Tire radar begins at the prevailing tire position, which is assumed to be the starting point. Block 1656 tests the presence of the INDEX pulse once per cycle and stores the Θ current at the "offset" location. If INDEX occurs, block 1657 stores the location in memory.

Block 1653 waits until the RPG is zero. When this condition

II

31 7281 6, block 1659 sets the pulse rate to 74 kHz and repeats the INDEX test in blocks 1660 and 1661 and waits until the RPG is number one in block 1662. The new pulse rate is selected in block 1663.

Once the complete RPG episode has elapsed, the data becomes ready for the flip-flop set and 1664 waits for this state. When the data is complete, the eight steady-state voltages generated by the integrators are converted at block 1665 and stored in the data memory as raw data. The position of the ring is incremented and tested with respect to the last data point in block 1666. The program continues to require data by jumping to retry location B. Once all points have been digitized and stored, the data is aligned in block 1667 so that the information associated with the INDEX point is at the beginning of the data block.

Although only one exemplary embodiment and only a few variations thereof are set forth in detail herein, it will be apparent to those skilled in the art that many modifications and variations can be made without departing from the novel and advantageous features of the present invention. Accordingly, it is intended that all such modifications and variations be included within the scope of this invention as defined by the appended claims.

Claims (7)

Apparatus for grinding and non-destructive testing of a ground tire cord or body, characterized by mutual resistance, circular rings (108r 110) arranged to abut closely against the corresponding annular inner edges of the tire, tire inflating means ( 102), which are arranged for inflating the tire after it has engaged the said rings, abrasive means (200) for rotating the inflated tire and for grinding the outer tread wall surfaces to form a tread wall surface which is substantially uniform exterior of the emerging tire cord or body, ultrasonic test means (210, 500, 502) disposed both inside and outside the tire deck (112) for measuring the relative strength of ultrasonic acoustic signals transmitted through various noise arrays . of said ground tread wall as it rotates so as to non-destructively test the ground tire cord or body. 2. Apparatus according to claim 1, characterized in that the ultrasonic test means comprise at least one ultrasonic acoustic transmitter (500, 502) and controllable mounting means (1504, 1502, 1500) of the transmitter, which are mechanically mounted in the rear ring for the transmitter. pulling the transmitter radially in the direction of the center of the circular rings during mounting and disassembly of the tire between the respective rings and for removing the transmitter radially away from the center of the circular rings in the direction of the tire runway wall during testing. An apparatus according to claim 2, characterized in that said adjustable mounting means (1504, 1502, 1500) of the transmitter, when in their radially retracted position, are also displaced axially with respect to at least one of said rings. (108, 110) to thereby facilitate the tire assembly operation without damaging said transmitter.