NZ205827A - Ultrasonic tyre testing:signals of different frequency at each test site - Google Patents

Description

Priority Dale(s;: . U9. £.1% Complete Specification Filed: P.V Class: Q.QJA.3S./.QU,.. trniiiw Publication Date: P.O. Journal, No: .. jr 205827 <Jnder tU< r .:j;ons of Regtfa ;lat!cn 23 (i; ie J i Specification has been ante-dated' to — 19d32_,' SE^ 1983 NEW ZEALAND PATENTS ACT, 1953 No.: Divided from New Zealand Patent'"Specification 193066 Date: 7 March 1980 COMPLETE SPECIFICATION METHOD AND APPARATUS FOR NON-DESTRUCTIVE INSPECTION OF TIRES £/We' BANDAG INCORPORATED, a corporation organized under the laws of the State of Iowa, U.S.A., Bandag Centre, Muscatine, State of Iowa, 52761, United States of America, hereby declare the invention for which I / we pray that a patent may be granted to nj^/us, and the method by which it is to be performed, to be particularly described in and by the following statement:- 2 058 27 This invention is generally directed to methods and apparatus for non-destructive inspection of rubber tires. Such inspection techniques may also be combined with conventional tire buffing operations in accordance with this invention.

There has long been an urgent need for ■cost effective, efficient, non-destructive inspection (NDI) of rubber tire casings. There are obvious safety benefits to be had by such techniques if they can be efficiently and rapidly practiced. There are also potential economic benefits. For example, during tire retreading operations, a defective tire carcass can be discarded before wasting further expenditures of time and money if it can be accurately, efficiently and cuicklv detected.

In fact, the need for improved NDI methods and apparatus relating to the testing of tire casings is so great that the U.S. Army Materials and Mechanics Research Center has sponsored special symposia devoted entirely to this subject in 1973, 1974, 1976 and 1978. The proceedings of the first three of these symposia have now been published and are available from the U.S. National Technical Information Service. They each include a corr.plete chapter on ultrasonic tire testing as well as other chapters devoted to different tire testing procedures (e.g. holographic, infrared and X-ray). There are also 2 0 5 8 2 many prior art patents relating generally to the use of ultrasonic waves to non-destructively test pneumatic tire casings. For example: U.S. Patent No. 2,345,679 - Linse (1944) 2,378,237 - Morris (1945) 3,336,794 - Wvsoczanski et al (1967) 3,604,249 - Wilson (1971) 3,815,407 - Lavery (1974) 3,882,717 - McCauley (1975) 4,059,989 - Halsey (1977) There are also several prior art patents relating to mechanical structures for chucking or otherwise physically handling pneumatic tire 15 casings during various types of non-destructive testing or manufacturing processes. For example: U.S. Patent No. 2,695,520 - Karsai (1954) 3,550,443 - Sherkin (1970) 3,948,094 - Honlinter (1976) 20 " " " 4 ,023 ,407 - Vanderzee ' (1977) Although a wide variety of nondestructive ultrasonic tests have been performed on tires in the past as shown by these prior art patents, they have each suffered serious 25 deficiencies and have failed to achieve widespread acceptance in commercial practice. Some of the prior art approaches have required a liquid ■ coupling medium on one or both sides of the tire wall under test. Some prior testing procedures 30 use a so-called "pulse-echo" approach which gives rise to a rather complex pattern of echoes due to r ^ 205827 normal internal tire structures as well as for abnormal structures. Many have used relatively low frequencies (e.g. 25 kHz) with resulting severe interference from normal ambient acoustic sources while others have used extremely high 5 frequencies (e.g. 2 jnHz) with resulting rapid ^ signal attenuation. Some have used continuous ultrasonic waves resulting in a confusing pattern of standing waves and the like while others have . looked for envelope "peaks in the received acoustic 10 waves. Others have used individual pulses of ^ acoustic signals for each tire testing site. In ! some cases the peak received envelope magnitude has'been used to reach final data values. Some have also attempted to test an inflated tire 15 carcass (but sometimes causing the acoustic signals to pass through two tire walls so as to keep all transducers external to the tire) although most have attempted to test a non-inflated tire carcass. There may have been other 20 techniques as well.

Prior tire chucking mechanisms in general have included axially movable tire mounting rims for quickly mounting and inflating a test tire. Prior NDI machines have located an ultrasonic ^ _ 25 transmitter inside a rotatable inflated tire, albeit such have been only fixed or manually adjustable mounting arrangements. Other NDI machines have included articulated transmitter mounting arrangement in conjunction with a spread-30 open non-inflated test tire. However, there has not yet been a commercially viable mechanical i i - 205827 arrangement for quickly positioning ultrasonic transducers about an inflated test tire wall while at the same time facilitating quick tire mounting/de-mounting procedures and also protecting the transducers from physical harm.

-It has now been discovered that these earlier attempts at ultrasonic non-destructive inspection of tire casings can be considerably improved and made more commercially viable.

For example, it has been discovered that a pulse or burst transmission mode may be used to reduce standing waves or unwanted reverberation effects within the tire. Each burst comprises only a few (e.g. 100) cycles of acoustic signals providing a very low overall duty cycle and 15 extremely efficient transducer operation. At the same time, it has been discovered that the envelope of received acoustic signals may be altered by internal reverberation, standing wave, wave cancellation or other irrelevant wave effects 20 after the initial portion or rising edae of each burst is received. In New Zealand Patent Specification 193066 from which the patent has been diyided, the .received acoustic signals are passed through a gated receiver circuit such that only those signals 25 within the initial portion of each burst are utilized.

I Still further improvements may be possible in some circumstances by averaging readings taken at different frequencies thereby 30 avoiding some possible adverse standing wave 205827 pattern effects and the like. Furthermore, nonlinear analog to digital conversion techniques may be used to assist in recovering usable data.

Accordingly the present invention consists 5 in a non-destructive tire testing apparatus for ultrasonically inspecting the wall or casing of a rubber tire, said apparatus comprising: an ultrasonic acoustic transmitter and an ultrasonic acoustic receiver mounted on opposite 10 sides of said tire wall, and electrical circuits connected to said transmitter and receiver for causing successive ultrasonic signals of different frequency to be transmitted through said tire wall at each of 15 successive testing sites and for combining measurements of received ultrasonic signals made at each site. 205827 A method and apparatus for non-destructive inspection of tires will now be described with reference to the accompanying drawings. Many features are described which are not all claimed in the present specification. These are claimed in the applicant's New Zealand Patent Specifications 193066, ^ioSlX? ,2o , ^ and In the drawings: 8 205827 FIGURES 1 and 2 are perspective views of a combined NDI/buffer machine constructed in accordance with this invention; FIGURE 3 is a block diagram of the invention shown 5 in FIGURES 1 and 2; FIGURE 4 is a. block diagram of the ultrasonic NDI circuits which may be used in the NDI/buffer machine of FIGURES 1-3 or in a machine having only NDI capabilities; FIGURE 5 includes'a schematic showing of a tire 10 wall section, acoustic transmitters and receivers and of the pre-amplif ier and multiplexing circuitry shown in FIGURE 4; ' FIGURE 6 is a detailed circuit diagram of the preamplifier shown in FIGURE 5; ■ 15 ■ FIGURE 7 is a detailed circuit diagram of a repre sentative one of the signal processing channels shown in FIGURE 4; FIGURES 8a and 8b comprise a detailed circuit diagram of the system interface shown in FIGURE 4; 2 0 FIGURE 9 is a detailed circuit diagram of the CPU or central processing unit shown in FIGURE 4; FIGURE 10 is a detailed circuit diagram of the display interface shown in FIGURE .4; FIGURE 11 is a schematic depiction of several repre— 25 sentative wave forms useful in explaining the operation of the circuits shown in FIGURES 4-10; FIGURE 12 is a cross-sectional view of a collimater/ impedance matching device used in each of the receiving transducers; 3 0 FIGURES 13 and 14 are tracings of CRT outputs ob tained by non-destructively inspecting a buffed tire carcass in accordance with this invention; FIGURE 15 is a detailed cut-away view of the fixed spindle and transmitter mounting arrangement of the embodi-3 5 . ment shown in FIGURES 1 ana 2; . .

FIGURES 16 and 17 are flow diagrams of a suitable control program for use with the CPU of FIGURES 4-10. " 205827 FIGURE 18 is a circuit diagram of another circuit for generating the AGC amplifier and integrator channels; and FIGURES 19, 20a and 20b illustrate a program 5 sequence which searches for air leaks and then searches for separations in two eight channel groups.

Referring to FIGURES l.and 2, two perspective views of the presently preferred exemplary combined tire buffer and NDI machine are shown. As will be apparent, 10 the NDI features of the machine may be provided, if desired, without including the tire buffing capability.

The major mechanical components' of the machine are mounted to an open frame 100 having a fixed spindle 102 and an axially movable spindle 104 opposingly aligned 15 along horizontal axis 106. Conventional circular tire mounting rings or flanges 108 and 110 are attached to the outer rotatable ends of spindles 102 and 10 4 for mounting an inflated tire 122, therebetween. A conventional pneumatically operated tire lift mechanism 114 20 is conveniently provided so as to assist the human operator in lifting and swinging a tire into and out of place between rings 108 and 110 during tire mounting and demounting operations.

Ring 108, and hence tire 112, is driven by 25 a two horsepower d.c. motor 116 through reducing gears 118. A tire surface speed of approximately 6 00 feet per minute is preferred for buffing operations while a much lower speed of approximately 4 0 feet per minute is preferred for NDI operations. Spindle 104, and 30 hence ring 110, is axially extended and' retracted by pneumatic cylinder 120. During tire mounting operations, ring 110 is retracted by cylinder 120 so as to permit the tire 112 to be lifted into place on ring 108 by lift 114. Thereafter, ring 110 is ex-35 tended against the corresponding rim of tire 112 and 20582 ■the tire is inflated to a desired set point pressure by compressed air passed through the center of spindle 102.

A conventional rotating tire buffing rasp 200 is mounted on a vertical pedestal 202 situated on the backside of t-he machine as seen in .FIGURE 2. The rasp 200 is controlled via a conventional panel 204 to move laterally along a desired buffing path 2 06 and horizontally towards and away from the tire by conventional control mechanisms including a "joy stick" used to control pneumatic cylinder 2 03, lead screws and associated drive motors and the like. The buffer rasp 200 is rotated by a separate motor mounted on pedestal 202 . The buffer mechanism, per se, is of a conventional type as marketed by Bandag, Inc., e.g. Buffer Model No. 23A- An array of 16 ultrasonic acoustic receiving transducers 210 is disposed above and around the outer walls of tire 112. The receivers 210 preferably include a conicallv shaped collimator and/or focusing tube to help limit the field of view for each individual transducer to a relatively small and unique area across the tire wall. The receivers 210 may be conveniently potted either individually or in groups in a polyurethane foam or the like, to help mechanically fix the receivers in their respective desired positions, to help protect the receivers and to help isolate the receivers from sourious ambient acoustic signals. The array of receivers 210 is radially adjusted into operative position by an air cylinder 212 having a coupled * n- 205827 hydraulic control cylinder so as to define a radially .extended operative position for the receivers 210.

A block diagram of the combined tire 5 buffer/NDI machine and its associated electrical and pneumatic circuits is shown in FIGURE 3. The electrical motor and pneumatic cylinder controls 300 are of entirely conventional design and thus not shown in detail. Operator inputs depicted at 10 the left of FIGURE 3 are made directly or indirectly by the operator via conventional electrical switches, | relays, aix valves and/or liquid control valves.

In operation, a tire is placed on lift 114 and raised into position between the rings 108 and 110 -5 Preferably, a predetermined index position on the tire is aligned with a physical index position or. flange 108. Thereafter, the chucking apparatus is engaged by causing flange 110 to move into the tire 112 so as to pinch the tire beads together in preparation 20 for tire inflation. The tire is then inflated to a desired.set point pressure. As will be explained in more detail below, the flange 108 is spring-loaded such that during chuck engagement and tire inflation, it is caused to move axially outwardly against .25 the spring-loading (e.g. by approximately 2 inches) . This facilitates the tire inflation process and simultaneously uncovers an ultrasonic transmitter located within the tire from a relatively orotected position so that it may subsequently be extended into 30 an operative posiiicn under ths arrav of receivers 210. An rnnerlocK switch activated by air oressurs anc/or by the physical movement of flange 108 may - ' 12 - 205827 be used to prevent any premature extension ojl Lhe transmitter before it is uncovered from its protected position.

In the buffing mode, the transmitter, need 5 not be extended.. The-buffing rasp drive motors are conventionally activated, and controlled. (e.g. with a "joy stick" and conventional push button controls) to buff the tire tread surface.as desired. Although it may not be required, it is 10 presently preferred to have the tire buffed to a substantially uniform outer treaawall surface before NDI operations are performed. Such buffing ' is believed to avoid possible spurious indications of defects caused.by normal tread patterns ana/or -5 by uneven wear about the tire surface. when the. operator selects the NDI mode of operation, an ultrasonic transmitter located inside the inflated tire 112 is extended into operative position and the array of receivers 210 is lowered into operative position by respectively associated pneumatic cylinders. The same 2-horsepower d.c. motor which drives the tire at approximately 600 surface feet per minute during buffing operations may be reduced in speed by conventional electrical circuits so as to drive the tire at approximately 4 0 surface feet per minute during the NDI mode. After the tire motion has reached a steady state, the operator may activate the scan request input switch to the ultrasonic NDI circuits 30 2. Thereafter the walls of tire 112 will be ultrascnically inspected curing one or more complete tire revolutions no crocuce a cisDlav 304 v/nich can be humanly interpreted directly or 20582 Indirectly to reveal the condition of the tire (e.g. satisfactory for further buffing and retreading, doubtful or unsatisfactory). If questionable condition Is Indicated, the tire may be discarded or may be additionally buffed and retested.

The ultrasonic NDI circuits 302 are shown In greater detail at FIGURES 4-10. As shown In FIGURE 4, the outputs from the 16 ultrasonic receivers 210 are amplified and multiplexed onto eight signal processing channels A-H by circuits 402 which are shown In greater detail In FIGURE 5. Each signal processing channel then provides AGC amplification, rectification, integration and ana log-to-dig ItaI conversion with the signal processing circuitry 404. A representatIve channel of such processing circuitry Is shown In detail at FIGURE 7. The resulting digitized outputs are presented to a conventional eight bit data bus 406 which Is Interconnected to a conventional micro-computer CPU (e.g. an 8080 type of eight bit computer) 408. The CPU 408 is also connected via a conventional address bus 410 and data bus 406 to a data memory 412, to a programmable read-only memory (PROM) 414 and to a system Interface circuit 416 which is shown in detail at FIGURE 8. A display Interface 418 (shown In detail at FIGURE 10) Is directly connected to the data memory banks 412 to provide a CRT type of oscilloscope display.

The system Interface 416 provides the necessary gating and other control signals to the signal processing circuitry 404 and also provides tire surfaces or too far away from these surfaces. In the preferred exemplary embodiment, transmitting crystals 500 and 502 are approximately two Inches from the inner tire wall surfaces although this optimum distance of separation may be varied by a considerable amount (e.g. plus or minus approximately one inch).

The arrayed receiving transducers 210 are located about an arc generally corresponding to the outside shape of the tire wall. Here again, it has been found that acceptable operation does not result if the receivers are too close or too far away from the outer tire walls. Preferably, the receivers are no closer than approximately 1 Inch to the outer tire surface but are preferably within 5.5 to 8.5 inches of the opposingly situated transmitting crystal. The receiving transducers 210 preferably each employ a conically shaped collimator and/or focusing tube as shown in detail at FIGURE 12. These tubes are preferably machined from polyvinyl chloride plastic material and also help to match the impedance of the actual transducer crystal surface to the surrounding ambient air acoustic impedance.

A moderately high ultrasonic frequency is employed so as to help avoid Interference from spurious ambient acoustic signals and to obtain increased resolution by using shorter wavelength acoustic signals while at the same time avoiding ultra-high frequence acoustic signals and the problems associated therewith. Frequencies above 40 kHz are desirable with 75 kHz 20582 being chosen as the presently preferred optimum frequency. Ultrasonic transducing crystals . operating at 75 kHz are conventionally available. For example, receiving crystals are available as the MK-111 transducer from Massa Corporation, Windom, Massachusetts, having the following specifications: Frequency of Maximum Impedance (fm) Impedance at fm. (min) Receiving Sensitivity (O.C.) at Frequency of Max Output Db re 1 Volt/microbar Transmitting Sensitivity Db re 1 microbar at 1 ft./10mw Maximum Power Input Directivity Temperature Stability 75 kHz±3 kHz Impedance at £m. (min) 6K Ohms Receiving Sensitivity (O.C.) at Frequency of Max Output -70DB min. -12DB Min.

OMW —10DB Max. at 90° Total Angle % Change in Frequency -30°F to +150°F Capacitance 120±20%PF A suitable transmitting crystal tuned to approxi-25 mately 75 kHz is available from Ametek/Straza, California under No. 8-6A016853.

The electrical leads from each of the transducers 210 are preferably connected through coaxial cables 506 to their respectively associated 30 pre-amplifiers 508. The outputs from each of the 16 amplifiers 508 are connected to an eight pole double throw electronic switch comprising Signetics SD5000 integrated circuits, controlled by the HIGH CHAN multiplexing signal provided by system interface 20582 _ 17 - 416. The eight resulting multiplexed output channels are connected through transistor buffer amplifiers to signal processing channels A-H. Accordingly, in the absence of a HIGH CHAN multiplex signal, the outputs from the first eight preamplifiers 508 are coupled to respectively corresponding signal processing channels A-H. However, when the HIGH CHAN multiplexing signal is present, the outputs from the last eight of the pre-amplifiers 508 are connected to respectively corresponding signal processing channels A-H.

The circuitry of each pre-amplifier 508 is shown in more detail at FIGURE 6. It includes a first transistorized stage having a gain of approximately 150.followed by a cascaded integrated circuit amplifier having a gain factor of approximately 11- The signal processing circuits 404 for each of channels A-H are identical. Accordingly, only the circuitry for channel A is shown in FIGURE 7. The waveforms shown in FIGURE 11 will be useful in understanding the operation of the circuitry in FIGURE 7.

The generation of a pulsed or bursted ultrasonic waveform for driving the transmitters 500 and 502 will be described later. However, by reference to FIGURE 11, it may be seen that each transmitter is driven to provide at least one approximately 50 cycle burst of 75 kHz acoustic output signals each time an RPGX trigger pulse occurs (e.g. 1,024 times per tire revolution). After a transmission delay, which will depend upon the separation between transmitter and receiver and the characteristics of the intervening 5 8 2 -]S- ambient air and tire rubber, the transmitted acoustic signals.are received. The received and transduced acoustic signals: may have a complex amplitude ■envelope (rather than the well-behaved one,shown. in FIGURE 11) depending upon the type of. multiple reflections, internal reverberations, wave cancellations, and/or other peculiar wave effects which take place along the'transmission path. Accordingly, it is only the leading edge or initial portion- of each 10 such ultrasonic pulse or burst (e.c. where the- amplitude envelope is initially increasing) . that provides the best and most accurate indication of the.transmission path quality (i.e. its included tire structural defects). Accordingly, the signal processing 15 circuitry shown in FIGURE 7 is adapted to effectively utilize only such initial or leading edge portions of each burst of ultrasonic.signals. In one embodiment, data for each tire measurement area is obtained by averaging measurements taken at different respec-20 tive acoustic frequencies.

As explained in U.S. Patent No. 3,8 8 2,717, it is necessary to provide automatic gain control amplification of through-transmission ultrasonic test signals to compensate for different average 25 tire casing thicknesses. This earlier patented system had but a single signal processing channel with AGC employed to compensate for differences in average tire casing thicknesses over the cross-section of a given tire. However, it has been dis-30 covered that automatic gain controlled amplification must also be included in each of the plural testing 2 0 5827 channels of this invention so as to compensate for differences In average tire casing thickness from tire-to-tIre.

Accordingly, an AGC amplifier 700 (e.g. Integrated circuit MCI352) Is Included within channel A as.shown In FIGURE 7. The ultrasonic signals passing through channel A are fed back to pin 10 of the AGC amplifier 700 and Input to a relatively long time constant (e.g. 10 seconds) RC circuit 702 connected to pin 9 of amplifier 700. Accordingly, the average of signals passing through the channel over the last several seconds (during the Included periods that the amplifier is enabled) Is compared to a constant reference AGC bias presented at pin 6 so as to maintain a substantially constant average output level at pin 7 over the RC time constant period. Amplifier 700 in the preferred exemplary embodiment has a gain which may vary automatically between a factor of 1 and 1000.

Amplifiers 704 and 706 are connected In cascade within channel A and each provide a gain factor of approximately 2. Additionally, amplifier 706 has diodes 708 and 710 connected so as to effect a full wave rectification of Its output signals as presented to the FET gate 712.

Referring back to FIGURE 11, an integrate reset signal INTGRST is generated during the first transmission delay period for a given test tire position and presented to FET gate 714 (FIGURE 7) so as to discharge the Integration capacitor 716 connected across amplifier 718 (forming a Miller-type integrator). Furthermore, the AGC amplifier 700 Is enabled by 20582 the AGCEN signal at some point during each testing cycle so as to sample the received signals. The integrator enabling signal INTGEN Is timed so as to enable the FET switch 712 only during the initial portions or leading edge of the ultrasonic burst (e.g. approximately 130 microseconds or about the first 10 cycles of the 75 kHz burst). If desired, two or more received bursts at respective different frequencies may be sampled and the results Integrated together so as to effectively average measurements taken at different frequencies (and hence having different acoustic standing wave patterns).

Thereafter, the output of Integrator 718 is converted to a digital signal under program control by CPU 408 generating suitable analog DAC Inputs to comparator 720 and conversion gatTng signals C0NV to gate 722 which interfaces with one of the conventional data bus lines (in this case DB0. Such program controlled analog-to-dfgftal conversion Is conventional and Involves the CPU program controlled conversion of reference digital signals to reference analog DAC signals which are then successively compared in comparator 720 with the results of such comparisons being made available to the CPU via data bus lines and gates 722. By a process of successive comparisons to different known reference signals, the programmed CPU Is capable of determining a digital value corresponding to the Input Integrated analog value from amplifier 718.

This process Is of course repeated simultaneously in channels A-H and successively in each channel for each burst or group bursts of ultrasonic 205327 Referring now to FIGURE 8, the RPGX (1,024 pulses per revolution) and RPGY (1 pulse per revolution) signals from the rotary pulse generator are passed through tri-state buffers 800 to data bus lines DB0 and DB1 respectively in response to' the IN3 and Q4 addressing signals provided by the CPU. Other addressing outputs from the CPU are input to an output decoder 802 so as to provide signals OUT320.00 through QUT320.70 under appropriate program -control.

Just prior to scan cycle, the CPU is programmed to repetitively poll data bus line DB2 looking for a scan request signal SCANRQ generated by an operator manipulation of the scan request switch 804 which causes flip-flop 806 to be set at the next occurrence of OUT320.60.

Once a scan request has been detected by the CPU via data bus line DB2, the CPU is programmed to poll the RPGX and RPGY signals which are then presented on data bus lines DB0 and DB1 by address inputs IN3 and Q4. An acutal measurement cycle is not started until the second RPGY signal is detected so as to insure that the tire is running true at a substantially steady state speed and that the AGC circuits are operating properly. Thereafter, each occurrence of an RPGX signal detected by the CPU is programmed to cause the generation of an OUT320.10 signal. The OUT320.10 signal triggers one shot circuits 808 and 810 and also enables the latch 812 to accept the digital values presented on data bus lines D3.0 through DB4. 20582: Just prior to the generation of the first burst of ultrasonic waves at a given tire wall test site, the CPU generates 0UT320.70 which triggers reset one shot 822 and provides an Integrator reset signal INTGRST via addressable flip-flop 823 and NAND gate 825.

The 4 bit binary counters 814 and 816 are connected in cascade to count the 18.432 Mhz clock signals Input from the CPU board and to divide these clock pulses by a numerical factor represented by the contents of latch 812. The result Is an approximately 75 kHz clock signal, (both 74 kHz and 76 kHz frequencies are used successively In one embodiment with the two results averaged together) which Is used to trigger one shot 818 having an adjustable time period such that Its output can be adjusted to a substantially square wave 50% duty cycle signal. As shown in FIGURE 8, one shot 818 is controlled by a pulser enabling signal from the addressable flip-flop 819. Thus If desired (e.g. to listen for leaks), the ultrasonic transmitters may be selectively disabled by the CPU.

The approximately 75 kHz 50^ duty cycle signal Is then buffered through amplifier 820 and presented as square wave output MB (see FIGURE 11) to conventional transmitter driver amplifiers (providing approximately 200 volts peak-to-peak electrical output) which. In turn, cause a generally sinusoid type of 75 kHz acoustic output from the transmitter as shown In FIGURE 11.

U215Efi;.

This generation of the approximately 75 kHz output MB will continue until one shot 808 times out (e.g. approximately 1 millisecond). During that interval, a burst of ultrasonic acoustic signals is caused to emanate from one of the transmIttIng crystaIs.

The period of one shot 810 Is adjusted for a delay approximately equal to but slightly less than the transmission delay between acoustic transducers. The delayed output from one shot 810 resets the data ready flip-flop 828 and triggers the Integrate timing one shot 826 which produces the Integrate enable signal INTGEN. At the conclusion of the Integrate enable signal from one shot 826, the data ready flip-flop 828 is set to provide a data ready signal to the CPU via data bus line DB4. If more than one analog data value is to be combined at the output of the Integrator, the CPU is simply programmed to ignore the data ready signal until the requisite number of measurement cycles have been completed. Ultimately, however, the data ready signal Indicates to the CPU that 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 DAC from the digital-to-analog converter 830 under control of the digital data latched into latch 832 from the data bus lines by the addressing signal 0UT320.00 At the same time, the CPU is programmed to provide proper conversion gating signals CONV via the addressing inputs to gates 834, 836 and 838. /rc Mi. rV *2I5£: , 2.0582 The DAC may be a linear type 08 or a non-linear exponential type 76 or other known non-linear types of DAC circuits. The non-linear DAC-76 Is believed to Improve the effective sIgnal-to-nolse ratio for lower level signals.

The CPU is programmed so as to normally produce the multiplexing HIGH CHAN output by setting and resetting the addressable flip-flop 840 via the address lines A0-A2, 0UT320.30 In accordance with the data value then present on data line DBO. However, manual override switch 842 has been provided so that either the low channels 0-7 or high channels 8-15 may be manually forced via tri-state buffers 844 with outputs connected to the data bus lines DB6 and DB7.

The flow diagram for an examplary CPU control program is shown In FIGURES 16-17. Conventional powei—up, resetting and initialization steps are shown at block 1500. After the START entry point, the scan request flip-flop 806 (FIGURE 8) Is reset, the Integrators are disabled (via flip-flop 823, FIGURE 8), and the data memory circuits are disabled at block 1502. Thereafter polling loop 1504 is entered and maintained until a SCANRQ on DB2 is detected.

Once a scan request has been detected, the Indicator lamps are tested, the Integrators are enabled for normal operation (via flip-flop 823), the data memory Is enabled for access by the CPU (and conversely, the display Interface Is disabled from access to the data memory) at block 1506. The hIgh/Iow/normal switch 842 (FIGURE 8) Is also checked via 2058 r DB6 and DB7. If the low or normal mode fs Indicated, the HIGH CHAN multiplex signal Is maintained equal to zero via flip-flop 840.

Thereafter, polling loop 1508 Is entered to test for an RPGY transition. A similar polling loop 1510 Is subsequently entered to Issue at least one tire revolution before measurements are taken. Then a software counter 0 current 's se+ ~*"° zero and the LOOP1 testing subroutine (FIGURE 16) Is entered. As will now be explained tn more detail, the step within L00P1 are executed 1024 times to col lect and record 1024 data values in each of eight transducer channels corresponding to 1024 tire testing sites distributed over a whole 360° of tire rotation in each of the eight channels.

After entry of L00P1, the RPGX signal on DB0 is tested for a transition from 1 to 0 at loop 1600. Once this transition occurs, ail the Integrators are reset (via one shot 822, FIGURE 8), the latch 812 Ts set to produce a 74 kHz MB drive signal and the transducers are driven with a burst of 74 kHz MB drive signals via one shot 808 and a pulser enabling signal via flip-flop 819. Since one shot 810 is also triggered, the leading edge of the received burst Is gated and integrated in each channel.

While this test at 74 kHz is being performed, the CPU is In a wait loop 1602. Thereafter, latch 812 Is reset to produce a 76 kHz MB signal and the transmitters are again pulsed. The result is another gated integration of the leading edge of a received burst at 76 kHz. As soon as 20582 this second Integration Is completed, the data ready signal on DB4 Is detected at waiting loop 1604. After the analog data has thus been accumulated for two different frequencies at a given tire test site, the AGC circuits are keyed (to keep them actively sampling the channel signal level within the relevant RC time constant period) and a conventional analog-to-digital conversion routine is entered. This routine converts each Integrator output to a six bit digital value which is then stored in the data memory 412. The data for each channel Is stored in a separate section of the memory so that similar data points for each channel can be later addressed using the same lower order memory addressing signals.

The Qcurrent software counter Is thereafter incremented by one and L00P1 is re-entered unless data measurements at all 1024 tire test sites have already been taken.

After the first exit from L00P1, a pattern recognition subroutine may be entered, If desired, at block 1512, The pattern recognition results may then be tested at 1514 and 1516 to determine which of status Indicator lamps 846 (FIGURE 8) should be lighted. Alternatively, the pattern recognition steps may be skipped as shown by dotted line 1518 to flip the HIGH CHAN multiplex signal, If operation is In the normal mode. (If only high or low channel testing has been forced by switch 842, return can now be made to the START entry point). Thereafter, measurements are taken for the higher group of eight channels as should now be apparent. 2058 While LOOP1 in FIGURE 17 causes measurements at 7 4 kFIz and 7 6 kHz to be combined, it should also be apparent that block 1606 can be skipped if measurements at only a single frequency 5 are desired. Similarly, measurements at more than two frequencies can be combined if desired. Furthermore, the combination of plural data values can be initially made either in analog form (as in the ' exemplary embodiment) or in digital form as should 10 now be apparent.

As already discussed, the CPU may be programmed, if desired, to automatically analyze the .digitized data collected during a complete scanning cycle with pattern recognition algorithms and to 15. activate one of the indicator lamps 846 (e.g. representing acceptance, rejection or air leakage) via conventional lamp driving circuits 848 as controlled bv the contents of latch 850 which is filled from data bus lines DBS through D34 under control of the address generated OUT320.20 signal. Air leakage can be detected, for example, by performing a complete scanning and measurment cycle while disablinq the ultrasonic transmitters.

Detected increases in received signals are then 25 detected as leaks.

The central processing unit shown in FIGURE 9 is conventionally connected to decode the various address lines and provide addressing inputs already discussed with respect to the system inter-30 face shown in FIGURE 8. The CPU itself is a conven tional integrated circuit 8080 microprocessor having cata input and output lines D.'J through D7 which are 205821 connected to the data bus lines DBO through DB7 through conventional bi-directional bus driver circuits 900. Address lines AO through A9 and A13 are also directly connected through buffer amplifiers 902 to the system Interface, memory circuits, etc. Address lines A10, All and A12 are decoded In decoder 904 to provide addressing outputs QU through Q7. Similarly, addressing lines A14 and A15 are decoded together with the normal writing and data bus Input signals from the CPU In decoder circuitry 906 to provide I NO through IN3 and OUTO through 0UT3 addressing outputs. The normal data bus input CPU signal DBIN and the addressing lines 814 and 815 are also connected through gates 908 and 910 to conventionally provide a directional enabling Input to the bi-directional bus drivers 900. The approximately 19 Mhz clock 912 is also conventionally connected to the 8080 CPU. However, pin 12 of the 3G8224 Integrated circuit Is brought out to deliver an 18.432 Mhz clock to the frequency dividing circuits of the system Interface already discussed with respect to FIGURE 8.

The data memory circuits are provided by a conventional connection of 25 Integrated circuits of the 4045 type so as to provide 8,192 eight bit bytes or words of data storage capability.

The programmable read-only memories may be provided by three Integrated circuits of the 2708 type, each providing 1,024 bytes of programmed memory. 256 eight bit words of read-write memory are also 20582.7 preferably connected to the CPU as part of the programmable memory circuits. An Integrated circuit of the type 2111-1 may be used for this purpose.

The CRT display Interface Ts directly connected to the data memory board. Once an entire measure cycle has been completed (e.g. when the third RPGY signal has been detected after a scan request), there are 1,024 data values available for each of the 16 measurement channels representing the relative magnitudes of ultrasonic signals transmitted through the tire at 1,024 successive respectively correspond Ing positions about the tfre circumference within the area monitored by the receiving transducer for a given channel. This digital data may be converted to conventional video driving signals for a CRT and displayed as shown In FIGURES 13 and 14. AlternatIvely, the 8080 computer may be programmed to analyze (e.g. by pattern recognition algorithms) the available digital data and to activate appropriate ones of the indicator lamps 846 shown in FIGURE 8.

The display Interface shown In FIGURE 10 Is conventionally connected directly to the data memory 412 via memory data bus lines 1000, memory quadrant selection bus lines 1002, memory address bus lines 1004 and data latch strobe line 1006. The whole display can be selectively disabled or enabled as desired under CPU control via CPU addressing outputs A13, Q3, 0UT3 and AO via flip-flop 1008 and the associated Inverter and gates shown in FIGURE 10. In the preferred 205827 embodiment, the display Interface Is disabled whenever other parts of the system are accessing the data memory 412 so as to prevent possible simultaneous activation of the data memory circuits.

The display Interface Ts driven by a 11.445 MHz clock 1010. Its output drives counter 1012 which Is connected to divide the clock signals by a factor of 70. The first 64 counts of counter 1012 are used by comparator 1014 which also receives 6 bits of data (i.e. 64 different numerical values) from the addressed data memory location representing the magnitude of ultrasonic signals transmitted through a particular tire testing site. Thus the output from comparator 1014 on line 1016 will occur at a specific time within 64 clock periods correspondIng to the magnitude of the Input digital data via lines 1000. The clock pulse during data coincidence will cause flip-flop 1018 to transition momentarily and produce a video output pulse via gate 1020 having one display dot time width and spaced within its respectively corresponding channel time slot according to the magnitude of the recorded data. Flip-flop 1022 Is triggered by counter 1012 upon counting a 65th clock pulse and generates an Inter-channel separation blanking video pulse out of gate 1020. The counter 1012 then continues to count 5 more clock pulses before resetting Itself and starting another cycle using data from the next adjacent channel. 2 0 5 8 2 The 70th count from counter 1012 also drives a three bit channel counter 1024 which, through the 3-to-8,decoder 1026, successively addresses eight different sections of the .data memory corresponding 5 respectively to eight of the sixteen ultrasonic receiver channels. 'A selection between display of the higher or lower eight channels is made via switch 1028 .- At the end of a complete horizontal scan 10 line, 10 x 70 clock pulses (2 x 70 clock pulses are counted during horizontal retrace period) will have been counted by counters 1012-and 1024 and a carry pulse will go to the 12 bit counter 1029 so as to increment the addresses on line 1004 (via decoder ]_5 1030) for the ne>:t horizontal scan line. In the case of the usual interlaced CRT scanning raster, every other horizontal line will actually be skipped and picked up during the second horizontal seam raster as will be appreciated. The states of counters 1024 and 20 1029 provide all requisite timing information for conventionally generating the usual CRT horizontal synchronization, vertical synchronization and vertical and horizontal retrace blanking video signals at 1032.

The various video signals are conventionally 25 mixed in video amplifier 1034 and output to a CRT display.

Since there are 1024 data values in each channel but many fewer horizontal scan lines in the usual CRT raster, switch 1036 is provided so as to 30 select only the odd or even addresses for data values in a given channel. Thus the complete 350° of scanned tire surface, within a given channel, is displayed in an assigned time slot over 512 ver t ically-spaced horizontal scan lines. 2 0 5 8 2 As thus described, the data values for a given channel would be distributed within a vertical segment of the CRT display and displaced in a horizontal sense from a vertical base datum line in accordance with the stored data values. However, in the preferred embodiment, the CRT deflection yoke is rotated by 30° so that the final CRT display for a channel is presented horizontally•as shown in FIGURES 13 and 14.

As depicted in FIGURES 13 and 14, the signal traces in each individual channel are deflected upwardly to represent reduced ultrasonic signal magnitudes. Accordingly, in FIGURE 13, it can be seen ' that a defect has occurred in channels 12 and 13 at approximately 20° from the index marker. Similarly, a defect is shown in FIGURE 14 at channels 12, 13 and 14 at approximately 280°.

Although not shown in FIGURES 13 and 14, if a leak had been present, it would have been indicated by an increased signal magnitude which, in the representation of FIGURES 13 and 14, would have resulted in a downward deflection of the signal trace for the corresponding channel.

The tr.acing for channels 0 through 3 and 12-15 is caused by wire ends, transitions between various normal tire layers and a periodic pattern of remaining tire tread structures about the outer edges of the tire treadwall. The data actually shown in FIGURES 13 and 14 was taken using a linear DAC circuit in the analog-to-digital conversional process.

Greater detail of the fixed spindle 102 and of the associated transmitter mounting arrangement is shown in the cross-section of FIGURE 15. The transmitting crystals 500 and 502 are directed at 90° with - 33 _ 2 0 5 8 2 respect to one another from the face of a PVC mounting block 1500. The' block 1500 is, in turn, attached to a retractable rod 1502 connected to the piston of a pneumatic cylinder 1504.

As shown in FIGURE 15, the pneumatic cylinder 1504 has retracted the transmitting crystals 500 and 502 into 'a protected area defined by an annular plate 1506 attached to the tire mounting ring or flange 108. The tire mounting ring 108 is rotatably secured to the 10 fixed spindle 102 through ball-bearing assemblies 1508 and 1510. This rotatable connection is maintained airtight by rotating seal assembly 1512. The center of the spindle 102 is hollow so as to permit passage ■of pneumatic control line 1514 and of the transmitter 15 electrical leads therethrough.

The rotating ring 108 and its connected assembly is spring-loaded via spring 1517 to its axially extended position as shown in FIGURE 15. However, the ring 108 may be moved axially to the 20 position shown in dotted lines against the spring force. In the preferred exemplary embodiment, such motion begins to occur when the ring 108 has approximately 1500 lbs. (2 psi) of lateral force applied thereto. The sliding joint which permits such motion 25 is also maintained airtight by "0" ring 1516. In the exemplary embodiment no more than approximately two inches of axial movement are permitted before the spring force is sufficient to resist further movement even when the tire is inflated to approximately 15-18 30 psi.

When the ring 108 is axially moved to the left as shown by dotted lines in FIGURE 15 against the force of spring 1517, transmitters 500 and 502 are then exposed and the pneumatic cylinder 1504 can be m 205827 activated to extend the transmitter into the position shown by dotted lines in FIGURE 15 for an operative measurement cycle. Suitable interlocking switches activated by the internal pressure of the inflated 5 tire and/or by the physical axial position of ring 108 can be employed to insure that pneumatic cylinder 1504 is not erroneously extended and damaged while the transmitters 500 and 502 are still enclosed and protected by the flange 1506.

FIGURE 18 shows another circuit for generating the AGC amplifier and integrator channels. The circuit permits generation of INTGEN, AGCEN, INTGRST, and MET pulses from RPGX pulses at 1605 or simulated RPG pulses from addressable latch 1608 under 15 program control.

When the RPG simulator is enabled, 1608 output labeled 5 is a 50% duty cycle pulse train wh.ich is selected by multiplexor 1611 to trigger one-shots 1512 and 1613. One-shot 1512 is triggered by the 20 rising edge of the output of 1611 and times out in 300 ns. One-shot 1513 is triggered by the falling edge of 1611 and also times out in 300 ns.

The outputs of 1612 and 1613 are combined to trigger DELAY one-shot 1614 and MB one-shot 1615. The 25 generation of 75 kHz bursts by 1615 , 1620, .1621, 1622 and 1623 has been previously described. DELAY one-shot 1614 triggers INTEGRATE one-shot 1616 and resets DATA READY flip-flop 1617.

Flip-flop 1617 signals that the analog 30 outputs of the AGC amplifier/integrator channels are ready for digitizing. Flip-flop 1617 is only set while RPG is high.

Flip-flop 1517 triggers AGCEN flip-flop 1619 which is level shifted anc sent to the AGC ar»Dl if ier s . 205827 A delayed RPG signal appears at the output of flip-flop 1618 and it is used by the softv.-are for synchronizing to tire rotation.

When the.simulator is disabled, multiplexor 1611 sends the logical output of 1605 to one-shots 1612 and 1613. The input source for 1611 now comes from the tire-rotation generated RPGX pulses, and the generation of the required outputs, i.e., INTGEN, is accomplished by controlling multiplexor -1611 outputting pulses to one-shots 1612 and 1613.

The sequence of one-shot firings•follows the same patterns as described in the previous paragraphs when the RPG simulator is activated.

The DAC comprised of 1624 and 1625 generate an analog voltage used by the CPU for analog-to-digital conversion of the intergrated values of received signals.

Decoder 1609, flip-flop 1610, register 1627 and lamp driver 162S perform functions already described. Latch-decoder 1529 and display 1530 provide status information during program execution.

During air-leak detection, PULSEN generated by software at 1608 is low, thus inhibiting MB excitation pulses to the pulser unit by clearing one-shot 1620.

FIGURES 19, 20a and 20b illustrate a program sequence which searches for air leaks then searches for separations in two eight-channel groups.

Blocks 1631 and 1632 initialize states of the system and 1S33 selects the RPG simulator to trigger the one-shot timing elements. The RPG simulator switches alternately high and low at a 8 ms rate while the SCAN; F.Q flip-flop is tested in the loop 1534 and 20582 1635. The RPG simulator refreshes the AGC levels so when SCAN RQ becomes active, data acquisition for air leaks can begin immediately.

When SCAN RQ becomes active, the RPG unit is 5 selected in 1636 and data memory enabled in 1637. Subroutine GETDATA is called at 1638 and is detailed in FIGURES 20a and 20b. Next, PATTERN REC is called at 1639, and any air leaks present will be detected and the. AIR LEAK lamp will be turned on by 1640 and 10 1641.

Now the pulser is activated at 1642. Tests for KICKAH, LOCHAM only and normal scan are done at 1643, 1644 and 1645.

Subroutines GETDATA and PATTERN REC are 15 called at 1646. Blocks 1647, 1648, 1649 and 1650 test for for REJECT/ACCEPT status and decide whether to continue to scan the high channel group. GETDATA and PATTERN REC are called again at 1651 and the tire status is tested again by 1652 and the program returns 20 to CONTINUE via 1653, REJECT status, or 16 54, ACCEPT status.

FIGURES 20a and 20b detail the flow of subroutine GETDATA. The position counter, 6 CURRENT is set to zero at 1655. The tire scan begins at the 25 current tire position which is assumed to be the origin. 31ock 1656 tests for occurrence of the once-per-revolution INDEX pulse and stores 8 CURRENT at location OFFSET. If INDEX is present, the 1657 stores the location in memory.

Block 165S awaits until RPG is zero. When the condition is met, 1659 sets the pulsing frequency to 7 4 kHz and repeats the INDEX test at 1660 and 1661, and waits until R.PG is one at 166 2. A new pulsing frequency is selected at 1663. 37 " 205327 When a complete RPG cycle has elapsed, the DATA READY flip-flop will be set, and 1664 waits for this condition. When DATA READY is true, eight steady state voltages generated by each of the integrators 5 are converted by block 1665 and stored in data memory as raw data. The tire position is incremented and tested for the last data point at 1666. The program continues to acquire data by jumping to the reentry point B. When all points are digitized and stored, 10 the data is justified in memory by 1667 so the data associated to the INDEX point is at the start of the data block.

While only a few exemplary embodiments and only a few variations thereof have been explained in 15 detail, those in the art will appreciate that many modifications and variations may be made without departing from the novel and advantageous features of this invention. Accordingly, all such modifications and variations are intended to be included within the 20 scope of this invention as defined by the appended claims.

Claims (7)

1. - - 20 582 7
2. WHAT/?WE CLATW IS: -1 • A non-destructive tire testing apparatus for ultrasonically inspecting the wall or casing of a rubber tire, said apparatus 10 comprising: an ultrasonic acoustic transinitter and an ultrasonic acoustic receiver mounted on opposite sides of said tire wall, and electrical circuits connected to said 15 transmitter and receiver for causing successive ultrasonic signals of different frequency to ba transmitted through said tire wall at each of successive testing sites and for combining fe ^measurements of received ultrasonic signals made 0 20 at each site'
3. W>' „ ^ • v 2. A non-destructive tire testing apparatus as in claim 1 wherein said electrical circuits include means for generating repetitive bursts or pulses of ultrasonic signals and means 25 for synchronizing the time at which said successive bursts are generated with corresponding successive increments of relative movement between said tire wall and the ultrasonic transmitter and rece iver. 30 3. a, non-destructive tire testing apparatus as in claim 1 wherein said electrical circuits include gated receiver circuits for 2058Z7 amplifying,/rectifymg and integrating the ultrasonic^Jfelectr ical signals produced by said jo receiver during e^ch of said gated time intervals thereby providing a succession of said electrical 5 measurement signals having respective magnitudes representative of the relative strengths of a succession of received ultrasonic acoustic signals.
4. 4. A non-destructive tire testing 10 apparatus as in claim-^% wherein said gated receiver circuits include an automatic gain controlled amplifier connected to automatically M[rjr control its gain in accordance with the magnitude ipnTJ of. ultrasonic/el^ctr^ical signals received during 15 an earlier scan of the same or other substantially similar portions of said tire wall.
5. 5_ A non-destructive tire testing apparatus as in claim j wherein said ultrasonic signals have a frequency higher than • :- 40 KHz. 20
6. 6. A non-destructive tire testing apparatus as in claim 1 wherein said ultrasonic signals have a frequency of approximately 75 KHz. i
7. 7. A non-destructive tire testing apparatus as in claim wherein said electrical 25 circuits utilize substantially only an initial or leading edge portion of each received burst when the envelope of said receiving ultra son ic signals is increasing in magnitude with respect to time. By_Ws/their authorised Agents, A. J. PARI'v, & Spt^