• Save
DESIGN OF A RELIABLE REMOTE SURFACE DETECTOR BASED ON ULTRASONIC PULSE-TRANSIT TECHNIQUE TO DETECT UNEVEN & NON-SMOOTH POROUS SNOW SURFACES
Upcoming SlideShare
Loading in...5
×
 

DESIGN OF A RELIABLE REMOTE SURFACE DETECTOR BASED ON ULTRASONIC PULSE-TRANSIT TECHNIQUE TO DETECT UNEVEN & NON-SMOOTH POROUS SNOW SURFACES

on

  • 1,414 views

This paper discusses design control techniques used to design ultrasonic pulse echo method based remote surface detector to detect objects with porous, un-even, irregular surfaces like snow. For such ...

This paper discusses design control techniques used to design ultrasonic pulse echo method based remote surface detector to detect objects with porous, un-even, irregular surfaces like snow. For such surfaces there is ample possibility of scatter of reflected beam in multiple directions, absorption of ultrasonic energy within the pores and reflected energy too feeble to detect. In some cases the reflected beam may altogether miss the receiver resulting in no surface detection at all. Scatter and penetration of ultrasonic energy at such irregular porous uneven surfaces poses great difficulties in designing a reliable Remote Surface detector. In this paper, we have presented a comprehensive set of various design techniques to develop a highly reliable remote surface detector. Experimental investigations have resulted into a highly accurate and reliable snow surface detector device being used in deep Himalayan regions for snow avalanche forecasting equipment.

Statistics

Views

Total Views
1,414
Views on SlideShare
1,414
Embed Views
0

Actions

Likes
0
Downloads
0
Comments
0

0 Embeds 0

No embeds

Accessibility

Categories

Upload Details

Uploaded via as Adobe PDF

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment

DESIGN OF A RELIABLE REMOTE SURFACE DETECTOR BASED ON ULTRASONIC PULSE-TRANSIT TECHNIQUE TO DETECT UNEVEN & NON-SMOOTH POROUS SNOW SURFACES DESIGN OF A RELIABLE REMOTE SURFACE DETECTOR BASED ON ULTRASONIC PULSE-TRANSIT TECHNIQUE TO DETECT UNEVEN & NON-SMOOTH POROUS SNOW SURFACES Document Transcript

  • R. Attri Instrumentation Design Series (Snow Hydrology), Paper No. 2, September 20051 | Page Copyright © 2005 Raman K. AttriDESIGN OF A RELIABLE REMOTE SURFACE DETECTOR BASED ONULTRASONIC PULSE-TRANSIT TECHNIQUE TO DETECT UNEVEN & NON-SMOOTH POROUS SNOW SURFACESRAMAN K. ATTRIEX-SCIENTIST (GEO-SCIENTIFIC INSTRUMENTATION), CENTRAL SCIENTIFIC INSTRUMENTS ORGANIZATION INDIArkattri@rediffmail.comAbstract:This paper discusses design control techniques used to designultrasonic pulse echo method based remote surface detector to detectobjects with porous, un-even, irregular surfaces like snow. For suchsurfaces there is ample possibility of scatter of reflected beam inmultiple directions, absorption of ultrasonic energy within the poresand reflected energy too feeble to detect. In some cases the reflectedbeam may altogether miss the receiver resulting in no surfacedetection at all. Scatter and penetration of ultrasonic energy at suchirregular porous uneven surfaces poses great difficulties in designinga reliable Remote Surface detector. In this paper, we have presenteda comprehensive set of various design techniques to develop a highlyreliable remote surface detector. Experimental investigations haveresulted into a highly accurate and reliable snow surface detectordevice being used in deep Himalayan regions for snow avalancheforecasting equipment.Keywords: Ultrasonic, Snow Surface, range detection,remote sensing, object detection, porous surfaces,piezoelectric sensorsI. INTRODUCTIONRemote Surface Detection (RSD) using Ultrasonic PulseTransit (Pulse-echo) method is one of the popular methods fordetecting the target surface, distance estimation and 3-Dimaging [1]. Mostly such methods are used for a targetshaving solid, smooth and possibly flat surface. For suchsurfaces the chances of reflected energy back to the source arehigh and the target is detected accurately [6, 7]. There aremany applications where the same method of Ultrasonic Pulsetransit has been extended for detection of targets withirregular, non-smooth surfaces, uneven surfaces or poroussurfaces [21].Snow Surface detection is one of the very importantapplications of such ultrasonic RSDs to determine thethickness of the snow layers by finding how much the currentsurface level snow fall is above the ground level. This requiresdetection of snow surface detection distance from the sensor[2, 18]. This information is used for critical and man-kindsaving hydrological studies such as forecasting of snowavalanche, river run off water, glacier sliding, river waterlevels and related phenomenon in the mountain areas andplanes nearby [4]. The fresh snow is extremely porous non-smooth and irregular surface where using remote sensing ofsnow using ultrasonic beams have its own problems [3, 5].Such surface detection requires lot many considerations indesign of such ultrasonic RSDs [8]. Such critically irregularand non-smooth porous surface (e.g. Snow surface) causespenetration of wave of incident wave into the surface,absorption in the surface and scattering around which resultsin either missed reflected wave or a very low amplitude highlynoise ridden reflected signal [11, 20, 21].Author conducted range of experiments using different designoptions to develop highly accurate and reliable snow surfacedetector device. As a result of multiple experiments a set ofdesign control techniques evolved which proved very effectiveand likely to be applicable to other applications involvingsimilar porous surfaces like sand, chemical compound fills,crops fillers etc. In this paper, we will present the tested designtechniques found effective for improving sensitivity, range,near-field side lobe compensation and temperature-velocityultrasonic waves.II. ISSUES WITH PERFORMANCE OF POROUSSURFACE DETECTORSExperiments revealed a number of practical issues withconventional remote surface detectors (RSDs) were observeddue to characteristic behavior of the snow surface and itsporosity,. Literature reports a number of issues on variousaspects when of ultrasonic transducers when used for surfacedetection applications.The snow typically is highly crystallized but porous surface.At all frequencies of transducer wave, the losses due toabsorption are virtually unavoidable [2, 3]. Penetration ofultrasonic waves at high frequency inside the snow surface isone of the biggest problems which require selection of properultrasonic generating transducers [11, 20]. The higherfrequency causes penetration in the porous snow surface andcause lot of heating and absorption in the surface [7]. At lowerfrequency of the ultrasonic, the penetration and absorptionlosses are less but accuracy of measurement suffers. On theother hand at high frequency penetration is more but range ispoor [20]. Therefore the selection of right frequency ofultrasonic transducer is very important for effectivemeasurement.Temperature dependence of impendence and capacitance ofthe piezo-electric transducers is also an issue. For the existingapplication, the transducer is expected to work in -20o C and
  • R. Attri Instrumentation Design Series (Snow Hydrology), Paper No. 2, September 20052 | Page Copyright © 2005 Raman K. Attrirequires a temperature stabilized transducer [9, 19].Temperature dependence of sound velocity and change indirectional pattern of the sensor due to fluctuation intemperature is also a big problem [9]. Temperatures may bedown to -40oC in heavily snow bound area.Another impairing factor comes by virtue of typical ultrasonicbeam shape at the transducer. Near-filed interference by sidegrating lobes in the beam triggers the receivers for false near-field object detection [10]. These grating lobes induce a signalin the nearest receiver situated on the side of these lobes;hence it is taken as reflected signal by the system, thus givingwrong distance reading as if the target surface is very near tothe sensor hood [12, 13].Irregular snow surface shape plays a vital role on the receivedecho amplitude and shape. It is seen that a roughness of morethan 1/10 of wavelength impairs the coupling markedly andmakes ultrasonic beam to become diffused and scatter in alldirections [10]. Sometime the reflected echo is not received atall. Another issue created by irregular non-smooth surfaces isthat more the directive the beam is more would be theprobability of scatter. These directive beams when reflectedfrom non-even surfaces may altogether miss the receiving area[12]. Physical design of receiver with proper “looking area”receiver is critical to enhance the probability and hence thereliability of the detector [14]. The biggest issue we faced iswhat should be the area covered by the beam width on theporous surface to ensure that sufficient coverage of toughs andcrest is ensured with high probability of finding a flat areafrom where there is a good reflection. Another importantimpact of surface irregularity seen during the experimentsemerges in form of constructive interference when incidentbeam strikes the surface at such an angle that the scatteredbeams make high amplitude interference pattern at thereceiver sensor. This give rise to detectable echo amplitudeeven when surface was beyond range of the system. This isessentially a false surface somewhere “in-between”.All design improvement apart, the strength of transmittedpulse governs the rest of the design. Higher the voltage attransmitter, higher is the power consumption which is aconcern in such field battery operator devices. Experimentallyit has been found that increase in transmitted voltageamplitude do have some improvement of results, but thisimprovement get saturated after certain point and no amountof increase in transmitted voltage make any effect [1, 7, 11].Choosing an optimum transmitter voltage level is an importantaspect of the design; because this governs the worst caseamplitude received at receiver and hence defines the loop gainrequired for a reliable operation. This also defines the softwareand hardware thresholds [11].Piezoelectric transducers typically act as charging-dischargingcapacitors with sinusoidal responses and harmonic response topulse train. The number of pulses in a single burst (or echo) totrigger sound mechanical waves has direct impact on the shapeand amplitude of received wave envelop. This in turn impairsthe successful detection of reflected wave. Using too lessnumber of pulses in the burst makes it difficult for the detectorto detect received echo successfully [7].During experiments, it was seen that the amplitude of thedetected received signal was too feeble to detect. Anattenuation of around 100db after reflection from such non-smooth surfaces located at around 4 meters from the sensorwas observed. This reflected echo need to be amplified to asufficiently detectable level [22]. Further the increase inamplification of received echo will only increase the noisebeyond the threshold level.The low amplitude reflected echo is further over-ridden inwhite noise. Noise amplitude is more than the signalamplitude so system with larger signal-to-noise ratio has to bedesigned. Inherent residual noise of the piezoelectrictransducers itself is great source of problem [22]. Oneimportant point observed is that in spite of the accuratefiltering of noise, the residual noise is always present at theinput of the sensors even when they are not receiving. Thereflected echo waveform gets added into it and if SNR ratio isnot controlled correctly, it makes detector almost impossibleto detect and filter the noise ridden echo [8].III. DESIGN TECHNIQUES & RESULTSWe carried out multiple design improvements throughexperiments. These improvements results in very reliable andsensitive remote surface detector which was testedsuccessfully on a porous snow surface. Following are themajor design changes.Physical and mechanical design improvementsFrom the experiments we saw that due to wavelengthinvolved, the accuracy of measurement is high at frequencysuch as 50 KHz, but penetration of ultrasonic waves is alsovery high thereby causing absorption losses. Also range ofmeasurement is too less at such high ultrasonic frequencies.The use of lower frequencies in the range of 20KHZ does havesome positive effects of increasing the range, reducing thelosses due to scatter and absorption due to directional patternbut then the accuracy of the measurement suffers. Refer to fig[1] which compares the range, accuracy and penetration ofultrasonic beam from piezoelectric sensors at differentavailable frequencies. Penetration is maximum at 50 KHz andrange is best with 25 KHz [20, 21].Taking into consideration the range, accuracy and penetrationinside the snow, 40 KHz crystal gives the optimumperformance, which has been selected. The availability oflower frequency moisture proof completely sealedpiezoelectric crystals working in –30 to +50 degreetemperature range limited our options. For present application,we have selected MZT-40E7S-1 from Murata Corp Japan
  • R. Attri Instrumentation Design Series (Snow Hydrology), Paper No. 2, September 20053 | Page Copyright © 2005 Raman K. Attri[19]. These are sealed piezoelectric transducer operating at 40kHz, as shown in Fig [2] and are used in dual mode- asreceiver and transmitter.Fig [1]: Graph of values of range, penetration and relative accuracyof Piezo-electric crystals as function of frequencyBeam width plays an important role in such cases [12, 13]. Itwas observed that sharper beams have more probability ofscatter at snow particles and broader beam have better chancesof reaching the receiver even during worst case reflection [26].The transducer selected for this application has the broaderdirectional response with directivity in sound pressure level of100oas shown in Fig [3] below.Fig [2]: MA40E7S-1 Piezoelectric sensors (Source: Murata CorpJapan) [19]These are dual-use waterproof sensors working at 40 kHz andable to work down to -30 degree Celsius. Working atmaximum voltage of 100V, the sensor is able to give 9mmresolution in range detection. With symmetric directivity of75o, it generates a broader symmetric beam of 106 dB soundpressure level at 30cm. Although the range of the sensor isspecified as 3 meters, however, with design improvementsdiscussed in this paper, we were able to get a range of 4.5meter in combination with recursive software algorithm fornear-field and far–field compensations.Fig [3]: Directivity of MA40E7S-1 transmitting piezo-electrictransducer in terms of sound pressure level measured in dB at 30cmwith 40 KHz frequency. Typical directivity is 75osymmetric in alldirections. This results in a broader beam with broad front zone areaof the beam (Source: Murata Corp Japan) [19]Even if the transducer geometry gave us a broad beam, it wasvery important to ensure that beam strength density is high atall the point in the front zone of the beam and overall surfacecoverage area should be broader. In order to deal with the non-smooth surface of snow and to counteract the scatteringeffects, the directional pattern of the transmitted ultrasonicbeam has been modified using concepts of array theory.According to array theory, the total directional response of anarray is simply the sum of directional responses of all theindividual transmitting elements [15]. The geometry of theseindividual elements shape up the directional pattern, beamwidth and side lobes. The spacing between the individualtransmitting transducers in x-axis and y-axis both affects thedirectional response. The mounting geometry i.e. whether inrectangle, square or triangle drastically change the directionalpattern [13, 15].Geometry (i.e. x and y spacing of transducer elements) isfound by using basic array theory calculations followed byextensive experiments to fine tune the right mountinggeometry. Based on experiments a line geometry withtransmitting array of 3 x 1 (i.e. 3 transducers mounted in arow) having spacing between them equal to little less than thehalf wavelength, was chosen. It is shown in Fig [4]. This arraygave sufficiently dense and powerful broad beam to avoid anychances of missing the reflected beam on the receivertransducer and to compensate for the losses due to non-smoothness. This design effort increases the possibility ofgetting the scattered beam even from the non-smooth surface
  • R. Attri Instrumentation Design Series (Snow Hydrology), Paper No. 2, September 20054 | Page Copyright © 2005 Raman K. Attriof snow.Fig [4]: 3 x 1 Transmitting Transducer connected in a series array toprovide superimposed phase-shifted transmitting beam envelop forhigh probability of receptionThe loss of power distribution over the broader beam widthwas compensated by increasing the power supply to thetransducers from 30V DC to 60V DC [19]. With thisconfiguration, the beam width becomes broader and powerful,so it travels as broader beam rather than a straight-line thinbeam towards the surface. Optimum performance can beachieved by varying the geometrical parameters further aswell as frequency of transmission, if we have an option.It was found during experiments that to improve theprobability of ‘worst-case’ reflection to be received atreceiver, the receiver face area should be designed correctlywith respect to transmitter beam width. With experiments, wefound that 1:3 ratio of transmitter and receiver works best foran optimum but highly reliable performance. We used 1x 3transmitter transducers in an array and 3 x 3 receivertransducer arrays. Essentially for one transmitter, there arethree corresponding receivers to receive the signal. Thisessentially increased the probability of reception by 3 times.The receiver array response was also optimised by using therectangular geometry consisting of array configuration withseparate 3 x 3 receiver array isolated from transmitting arraymechanically as well as electrically, as shown in Fig [6].Wider receiver surface area gives the ample opportunity thatthe reflected beams at some angle will also get captured. TheTechniques really made the system very robust thus workedwell with almost all kinds of rough surfaces and particularlyproved suitable on snow surface.There are two options of connecting the array: one is puttingelements in parallel and one is putting the array in the series[15]. In our case we have chosen the later approach. It hasgiven a big design advantage. Putting transmitter array inseries, one shown in Fig [5] ensured that each of the threetransducers are activated with little delay in between, so resultis transmission of three envelops of bursts. This short delaymakes the right superimposition at the rough surface andchances of reflection from any one of the envelopes arestrengthened. This will be clearer from the Fig [6] whichshows the three transmitted envelops received at the receiver.Fig [5]: Array arrangements for Transmitter and Receiver. Receiverarray is a rectangular array of 3 x 3 and transmitter array is a linearray of 3 x 1.Fig [6]: Superimposed time shifted 3 transmitted wave frontenvelops. Due to capacitive effect, the three transmitters gettriggered with a short delay in between. This results in asuperimposed three envelops of pulses triggering in sequenceSimilarly the receiver has also been connected in series, asshown in Fig [7]. This gives a very big advantage that theoverall voltage received at the receiver section issuperimposition of the wave front arriving at each of thereceiver transducer. This strengthens the receiver signal andeven if only one transducer element has received the wavefront, it acts as right input signal.The receiver section in this case will receive the three echoessince there were three transmitted envelops. This ensure thateven if we are transmitting one envelop consisting of Nnumber of pulses, the receiver wave front will be asuperimposed wave front consisting of 3 x N number ofpulses. This increases the system responsiveness and lessens
  • R. Attri Instrumentation Design Series (Snow Hydrology), Paper No. 2, September 20055 | Page Copyright © 2005 Raman K. Attrithe chances of missing the reflection.Fig [7]: 3 X 3 rectangular receiver transducers array connected inseries array. The series connection ensures the summation of theenvelops received by each receiver transducer to create asuperimposed resultant signal at amplifier inputsB. Electronics Circuit Design improvementsIn pulse-transit method, only a short burst of pulses is senttoward the target. Using a single pulse is not effective in givensituation as there is good probability of single pulse distortion.Even when multiple pulses were used the inherent dampingfactor of the transducer at the transmitter side as well asreceiver side made these pulses to exhibit modulated envelopewith peak at centre of the envelope. The pulses at the extremesof this have very low amplitude [1]. For successful detectionof the received echo, it is very important that the amplitude ofat least one of the received pulse in the burst should be higherthan the set threshold. Refer to Figure [8], which shows whatis being transmitted.Experiments showed that 16 numbers of pulses were optimumfor detection of porous snow surface. We used an electronicscircuit to control transmission of 16 pulses. A single pulsetriggers the transmission section. Transmitter operated on 40kHz frequency generated by 40 kHz oscillator. An R-S flip-flop in conjunction with the 16 pulse counter was used totransmit only 16 pulses to the base of Transmitter Amplifier asshown in Fig [9]. The duty cycle of 25 % was chosen to givetolerance for the transition time of transistors and rise time ofultrasonic transducers.The primary response of the Remote Surface Detector dependsupon the amplitude of the transmitted signal. Greater theamplitude better shall be the reflected echo. Normallypiezoelectric ultrasonic transducers are operated at highvoltages as high as 100 V. Experimentally it were observedthat increase in transmitted amplitude have very minimaleffect on the response of the system. With increase in 10% oftransmitter amplitude voltage only 1% improvement in theresponse is achieved [11]. So transmitter voltage could not beraised beyond a threshold, if raised at all, the improvement ofresponse was minimal at the cost of heavy currentconsumption. 60V was chosen as optimal transmitter voltage.Fig [8]: A burst of 16 ideal square pulses of 60V triggered by outputamplifier results in modulated mechanical waves at the transmittertransducer. The received signal also exhibit similar modulatedcharacteristics.Fig [9]: Transmitter Section is a Class-C amplifier operating at 60V,driven by a switching transistor. An oscillator and RS flip-flopcombination logic ensures passage of 16 pulses to power amplifierand to series array of transducers.Output of the receiver piezoelectric sensor is as minute as 1mV in response to the 60 V transmitted signal. This reflectedecho is shadowed so severely by the noise of higher amplitudethat it was impossible to detect it without further signalsconditioning. The reflected echo is applied to the inputamplifier stage which server the dual purpose of impedancematching as well as initial gain. Refer to Fig [10] showing thereceiver circuit diagram in detail.
  • R. Attri Instrumentation Design Series (Snow Hydrology), Paper No. 2, September 20056 | Page Copyright © 2005 Raman K. AttriFig [10]: Block diagram of the receiver wave shaping circuit. Thecircuit amplifies the weak received signal, compare it with a nominalthreshold to eliminate the white noise and shape the resultant signalinto square wave pulsesInputs stage gain was designed at 200. The amplified signalwas still in mV range as shown in Fig. [11a] and was highlynoise ridden spanning over the entire spectrum. A high gainamplifier with an optimum gain of 50 was used for furtheramplification of the signal. The useful signal bandwidth isapproximately 50 kHz, which implies unity gain bandwidth ofapproximately 2500 kHz. A video amplifier with UGB of 120MHz has been used to amplify all the frequencies uniformly.The output signal of the order of 1 V was received but stillshadowed by the noise of higher amplitude. As shown in Fig[11b] the reflected echo is still not reproduced faithfully.Experiments show that the high amplitude noise is well aboveor well below centre frequency of 40 kHz. A precision filteramplifier was designed as narrow band pass filter with sharpcut-offs at 38 kHz and 42 kHz with centre frequency of 40kHz. The response of this filter stage is shown in the figure[11c].The resultant echo of very clear shape and amplitudewas received.The number of pulses received in this echo is three times thenumber of transmitted pulses (because of three transmitters).Actually the received echo consists of three envelopes shiftedslightly in phase and each envelope having 16 pulses. Theincreased number of pulses gives better chances of triggeringof next circuit and hence the measurement and hence bettersensitivity.One important point observed is that in spite of the accuratefiltering of noise, the residual white noise is always present atthe input of the sensors even when they are not receiving. Thisresidual noise gets amplified and appears at the output of thefinal filter amplifier and is present all the time. The reflectedecho waveform gets added into it. Generally cross convolutiontechniques are employed to remove such residual noise. Inpresent case alternative method of comparison with somefixed threshold has been used [23]. The average value ofresidual noise is checked out. A comparator circuit as shownin Fig [11c] is used to compare the filtered output to thisthreshold voltage. If the reflected echo above this threshold,comparator gives the output in the form of pulses crossing thethreshold mark. The comparator also wave-shape the bipolarsine signals into unipolar square pulses. One importantprecaution in this design is the setting of gain and thresholdvalue of the circuit. A trade-off point have been observedbeyond which accuracy may get impaired if we further lowerthe threshold and increase the amplification in order toincrease the system sensitivity.The result of using the comparator is that by setting thethreshold correctly, we can ensure to receive the trigger fromthe genuine echo signal. The effect of threshold setting can beseen in Fig [11d], which helps to detect only the meaningfulsignal above the threshold and hence discards the residualnoise.The pulses received are conditioned and wave-shaped to drivethe digital circuit, refer to fig [11e]. First pulse out ofincoming pulse stream pattern as shown in Fig [11f]) isextracted as only this pulse corresponds to arrival of the echoitself. First pulse extracted from transmitted stream starts atiming counter and this first pulse of incoming stream stopsthis counter. The counts in the counter are direct indication offlight time (time taken by the echo to travel to target andreturn back to the sensor). This time is used in computation ofdistance of target surface from the sensor.While testing the system, it was observed that the two streamsof the received pulses is obtained, first one with largeamplitude. The first stream comes immediately after thetransmission and other stream arrives late. The first one causesa false triggering of the receiver and causes the time counter tostop, as a result of which system presume there is an object inthe close proximity. Investigations showed that this is becauseof one of the two factors: First, the broader beam dispersion oftransmitting beam (typically 75o) may trigger the receivertransducers spaced marginally. Even when a shield is appliedbetween the transmitter and receiver pairs, there are goodchances of false triggering from the beam edges touching thereceiver reception area. The effect is shown in the Fig [13].Secondly this false trigger could be caused by the side gratinglobes of the beam. Although as per manufacturer data,directional pattern has no side lobes, but the effect oftemperature on changing directional pattern at -30 degree isnot known. Irrespective of the source of this interference(whether due to broader beam or due to side lobe), the countermeasure is same.
  • R. Attri Instrumentation Design Series (Snow Hydrology), Paper No. 2, September 20057 | Page Copyright © 2005 Raman K. AttriFig [11]: a) Input signal at Preamplifier Transistors with very low amplitude ridden in the mechanical noise b) Output signal from GainAmplifier which amplifies the weak signal as well as amplifies the background noise. The prominent signal level corresponds to reflected echo.First saturated signal corresponds to either side lobes or false trigger due to broader beam c) Output signal at Filter Amplifier output whichcleans the signal and removes the noise. The reflected echo signal is enhanced for digitization. d) Pulse stream at comparator which removes thebackground residual noise and digitizes the inputs signal peaks above the threshold voltage e) Pulse stream at Wave shaper Output removes thenegative side of the pulses and produces clean rectangular pulses from 0-5V f) Output of 16-to-1 pulse isolator circuit which produces one pulseagainst a stream of 16 pulses. This single pulse is used to stop the timing counter for measuring transit time
  • R. Attri Instrumentation Design Series (Snow Hydrology), Paper No. 2, September 20058 | Page Copyright © 2005 Raman K. AttriFig [12]: Near-field interference due to broader transmitter beam ordue to side grating lobes.As a counter-measure to stop false detection, we developed acircuit and software design to block first stream of pulses. TheFig [13] shows the circuit for blocking the first stream ofpulses till the total worst-case duration of first stream so thatthe sensor responds to actual echo only. The voltage level atinput of OR gate is made high till the expected time of firststeam. The expected time has been around 2300microseconds. A safe margin is given and pulses till the time2560 microsecond are blocked. After that any stop pulsearriving is passed through and the stop the counter. In softwareany reading below counts corresponding to this time arerejected and measurement is termed false and cycle is againrepeated. The downside to this implementation is that sensorcan not effectively detect any surface within 50cm. This iscalled blind range. However, this was acceptable in our case,since snow surface is never likely to touch the sensor hood.Fig [13]: Near-field compensation circuit to stop the system counterto avoid false triggering of receiver circuit due to broad beam patterfrom transmitter and / or side grating lobes.C. Compensation and improvements through SoftwarealgorithmDue to irregularities of the surface the reflected beam maymiss the receiver and in system counter may keep waiting forthe reflected echo. To bring the system out of wait mode,software checks for the maximum expected counts. If currentcounts crosses the maximum permissible counts, system takethe situation as missed beam and repeat the measurementcycle until it gets the counts in the permissible accurate range.After repeating the cycle for 3-4 times, if a valid reading is notavailable, the system indicate the message on the LCD that notarget surface found with-in range. This warning messageallows the operator to take corrective actions.There may be a kind of constructive interference patternedformed at receiver due to superimposed scattered beams. Thereason for this is not fully known, but these results in someinconsistent measurements. We postulated that it probably isdue to broadening of the beam and hence multiple reflectionsfrom multiple sides. In order to create the repeatability in thereadings and to remove such inconsistent measurements,through software algorithm the same measurement is repeatedthrice in succession and a comparison is made if all the threereadings are in close proximity. If they are in close proximity,the average of the distance measurement readings is taken andhence repeatable results are achieved. In case the readings arenot in close proximity, then software logic discard this as falsetriggering and repeats the cycle. This far-field compensationenhances the system performance [8].The velocity compensation has to be provided in the design ofthe system. To counter this dependence, an integrated solidstate temperature sensor has been used along-with itslinearized circuit, which reads the temperature with theresolution of 0.01ocentigrade [16, 17]. 8-bit ADC is used fordigital conversion of temperature. Extensive Softwarecomputations have been carried out to compensate the effectsof the temperature over the velocity of the sound.IV. EXPERIMENTAL SETUP OF REMOTESURFACE DETECTOR (RSD)A test setup was erected as shown in Fig [14] inside the coldchamber controlled at a specific temperature near 0 degreeCelsius. The fresh snow cover was collected and dispensedusing a snow dispenser over a non-movable platform of 25square meters. The test platform of dispensed/ scattered freshsnow inside the cold chamber gave close to realistic test setup.The care was taken not to make the fresh snow flattened onthe platform. The platform was basically a cuboids’ containerwith height of 10cm. The dispenser filled the container till thetop edge with substantial pores, irregularities and un-evenness.For measurement purposes, height of snow cover top surfacewas taken as the height of the platform edge i.e. 10 cm.The RSD was mounted on a rigid mast of 5 meter height withsensor facing downwards toward the platform filled with freshsnow. The typical mounting setup is shown in Fig [15]. TheRSD mast movement was controlled by a hydraulic system tolower its height vertically downwards towards the platform inthe increments of 1 cm. The physical distance of top surface ofthe snow cover on the platform from the sensor was calibratedin the beginning. The reading on distance of snow cover fromthe transducers as reported by RSD was stored along with the
  • R. Attri Instrumentation Design Series (Snow Hydrology), Paper No. 2, September 20059 | Page Copyright © 2005 Raman K. Attriphysical distance of the RSD from the snow cover.Fig [14]: Remote Surface Detector Mounting ArrangementV. DISCUSSION ON SYSTEM PERFORMANCEThe system has been designed to work not only for samplingbased measurements, but also for continuous round the clockfunctioning as well. The long-term stability of the readings ischecked by recording the distance shown by this systemcontinuously for 1 hour and averaging those out to a singlenumber was taken as the distance reported by RSD. The errorbetween RSD reading and actual reading is plotted along Y-axis in the Fig [15]. Actual distance as recorded physically isplotted against Y-axis.The accuracy of + 1 cm with the resolution of 1 cm isachieved within the range of 4.5 meters. On smooth flatsurfaces, system measure up to 5 meters with + 2 cm accuracy.System performance over all kind of porous and irregularsurfaces has been checked out and has been found veryreliable, repeatable and accurate within + 1 cm up to 4.5meters of range. Due to reinforced array technique andsoftware repeated measurements, optimum spacing andbroadening of the beam; the reliable measurement has beenachieved up to 4.5 meter. The wave is sometime missed andnot received back as echo if the senor is away from targetsurface more than 450 cm (far field). System software catchessuch events and repeats the measurements and average out thefalse detections. On the other hand near field error comes inpicture due to side lobe compensation and system does notdetect any surface when it is in proximity of 50 cm to thesensor face. The side lobe compensation circuit sacrifice nearfield 50 cm measurement range, which practically is neverused. The snow surface hardly reaches within 50 cm proximityof the sensor.Fig [15]: System performance diagram showing error in cm againstthe zero error axis at different points within the range. The systemrecords distances between 50cm and 450 cm accurately. As it movesaway from 450cm, the scattering effects comes into picture whichmay make the reflected beam to miss the receiver altogether. Theaccuracy is impaired in far field. Detection of surface is inhibited upto 50cm in the near field due to side lobe compensation (the captionson the points indicate the measured absolute distances in cm)VI. CONCLUSIONSThe system has specifically been used for snow depthmeasurement which is extremely important for modelling ofsnow avalanche forecast and other related studies. The freshsnow surface being the most porous surface and non-smooth,the design approach for snow surface would surely beapplicable to all the other materials. Because of this the systemcan be used for industrial, commercial, defence applications.The experimental results have been very favourable.Considerable amount of reflected signal is received througharrayed receivers, which is detected with the help of theelectronics design. The sensitivity, accuracy, long-termstability and range of the system have been enhanced usingthese design techniques. The surface distance more than 5meters can be fairly detected in spite of the non-even, rough,porous non-smooth graining of the surface which normallydoes give problems of scattering and absorption of energy into the material.ACKNOWLEDGEMENTSDr. B.K. Sharma, Head Of Dept, Geo-Scientific InstrumentsDivision, CSIO, Chandigarh
  • R. Attri Instrumentation Design Series (Snow Hydrology), Paper No. 2, September 200510 | Page Copyright © 2005 Raman K. AttriSwaranjit Singh, Senior Technical Officer, Geo-ScientificInstruments Division, CSIO, ChandigarhREFERENCES[1]. G.L Gooberman., Pulse Techniques, Ultrasonic Techniques inBiology and medicine, Illiffe books Ltd, London, 1967.[2]. S. Kumar et al, Snow Depth Senor, Proceeding of nationalsymposium on Sensors and Transducers, 1996.[3]. M. Mellor, Engineering properties of snow, Journal ofGlaciology, volume 19, pg 15-66, 1977.[4]. T.J. Yamazaki Kondo, T. Sakuraoka and T. Nakamura, A onedimensional model of evolution of snow cover characteristics, Journal ofGlaciology, Vol. 18, pg 22-26,1993,[5]. K. D. Hall, Remote sensing of Ice and Snow, London, Chapmanand Hall Publications, 1985,[6]. M.C. Combs; Jr. Goodwin, H. Perry, Adjustable ultrasonic levelmeasurement device, United States Patent 4221004, Aug 1978[7]. M. Krause, et. al. Comparison of Pulse-Echo-Methods for TestingConcrete, E-Journal of Non-destructive testing, Vol.1 No.10, October 1996,[8]. A. Hämäläinen and D. MacIsaac, Using Ultrasonic Sonar Rangers:Some Practical Problems And How To Overcome Them, Phys. Teach. Vol 40,pp 39, 2002.[9]. D. Zhang D, G.M. Crean, Non-linear acoustic properties of highpurity quartz as function of temperature, Proceedings of Development inAcoustic and Ultrasonic, IOP Physical Acoustic Group, Leeds, pp 219-224,Sept 1991[10]. D.S. Balantine et al, Acoustic Wave Sensors-theory, design andphysico-chemical applications, Academic Press, 1997[11]. D. Ensminger, Ultrasonic – the low and high intensity applications,Marcel Decker Inc, New York, 1973[12]. I Busch, E. Huxson, Ultrasonic, Encyclopedia of applied Physics,VCH Publishers, vol 1, pp 63-88, 1991[13]. Ultrasonic, Encyclopedia of Science and Technology, Vol 12, pp662-664, Mcgraw Hill, 1982.[14]. E.P. Papadakis, Physical acoustic principles and methods (W.P.Mason , ed.), vol 4 B, Academic, NY, 1968[15]. J.E. Hudson, Adaptive Array Principles, Peter Peregrinus , London1981[16]. P. Klonowski, AN-273: Use of the AD590 TemperatureTransducer in a Remote Sensing Application, Analog Devices ApplicationNotes, www.analog.com[17]. M. P. Timko, A two-terminal IC temperature transducer, IEEEJournal of Solid-State Circuits, vol. SC-11, 1976, pp. 784-788.[18]. SR-50 Sonic Distance Sensor, Campbell Inc, Canada,http://www.campbellsci.com/documents/lit/b_sr50.pdf[19]. Piezoelectric Ceramic Sensor (Piezoliote), Cat-P19-E9, MurataManufacturing Co. Ltd, Japan, http://www.murata.com/catalog/p19e.pdf[20]. M.G. Silk, Ultrasonic transducers for non-destructive testing,Adam hilger Ltd, Bristol, 1984[21]. L. Krautkramer and H. Krautkramer, Ultrasonic testing ofmaterials, Berlin, Springer, 1969.[22]. S Raman et al, Processing of Ultrasonic signal for transducercharacterization and for improving signal to noise ratio, Indian Journal ofTechnology, Vol 31, Nov 1993, pp 774-776[23]. J.L. Lawson and G.E. Uhlenbeck, Threshold Signals, Mcgraw Hill,1948 pp 211Author Details:Author is Global Learning and Training Consultantspecializing in the area of performance technology. Hisresearch and technical experience spans over 16 yearsof project management, product development andscientific research at leading MNC corporations. Heholds MBA in Operations Management, Executive MBA,Master degree in Technology and Bachelor degree inTechnology with specialization in Electronics andCommunication Engineering. He has earned numerousinternational certification awards - Certified ManagementConsultant (MSI USA/ MRA USA), Certified Six SigmaBlack Belt (ER USA), Certified Quality Director (ACIUSA), Certified Engineering Manager (SME USA),Certified Project Director (IAPPM USA), to name a few. In addition to this, he has60+ educational qualifications, credentials and certifications in his name. Hisinterests are in scientific product development, technical training, managementconsulting and performance technology.E-mail: rkattri@rediffmail.comWebsite: http://sites.google.com/site/ramankumarattriLinkedIn: http://www.linkedin.com/in/rkattri/Copyright InformationCopyrights © 2005 Raman K. Attri. Paper can be cited withappropriate references and credits to author. Copying andreproduction without permission is not allowed.