• Save
DESIGN STRATEGY OF SNOW DEPTH SENSOR BASED ON ULTRASONIC PULSE-TRANSIT TECHNIQUE FOR REMOTE MEASUREMENT OF SNOW COVER THICKNESS
Upcoming SlideShare
Loading in...5
×
 

Like this? Share it with your network

Share

DESIGN STRATEGY OF SNOW DEPTH SENSOR BASED ON ULTRASONIC PULSE-TRANSIT TECHNIQUE FOR REMOTE MEASUREMENT OF SNOW COVER THICKNESS

on

  • 1,378 views

Snow cover thickness is one of the important parameter used in forecasting models for snow-melt, snow run-off water, snow avalanche release and other snow hydrological changes. Ultrasonic pulse ...

Snow cover thickness is one of the important parameter used in forecasting models for snow-melt, snow run-off water, snow avalanche release and other snow hydrological changes. Ultrasonic pulse transit method is being used for such applications universally. Reflected echo coming after reflection from highly irregular and non-smooth porous surface, is very low amplitude noise ridden signal. This received echo signal has to be conditioned to remove signal and processed to increase its amplitude to make it sufficiently detectable and to increase the probability of receiving back the reflected echo. A special design of Snow Depth Sensor based on Ultrasonic Pulse transit Method to improve the sensitivity, response and reduction of losses has been worked out. This paper discusses the signal processing technique and electronics design approach for development of this Snow Depth Sensor.

Statistics

Views

Total Views
1,378
Views on SlideShare
1,377
Embed Views
1

Actions

Likes
0
Downloads
0
Comments
0

1 Embed 1

http://www.linkedin.com 1

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 STRATEGY OF SNOW DEPTH SENSOR BASED ON ULTRASONIC PULSE-TRANSIT TECHNIQUE FOR REMOTE MEASUREMENT OF SNOW COVER THICKNESS Document Transcript

  • 1. R. Attri Instrumentation Design Series (Snow Hydrology), Paper No. 3, August 1999Copyright © 1999 Raman K. Attri Page 1DESIGN STRATEGY OF SNOW DEPTH SENSOR BASED ONULTRASONIC PULSE-TRANSIT TECHNIQUE FOR REMOTEMEASUREMENT OF SNOW COVER THICKNESSRAMAN K. ATTRIEx-Scientist (Mechatronics & industrial automation) Central Scientific Instruments Organization INDIArkattri@rediffmail.comAbstract - Snow cover thickness is one of the importantparameter used in forecasting models for snow-melt, snowrun-off water, snow avalanche release and other snowhydrological changes. Ultrasonic pulse transit method is beingused for such applications universally. Reflected echo comingafter reflection from highly irregular and non-smooth poroussurface, is very low amplitude noise ridden signal. Thisreceived echo signal has to be conditioned to remove signal andprocessed to increase its amplitude to make it sufficientlydetectable and to increase the probability of receiving back thereflected echo. A special design of Snow Depth Sensor based onUltrasonic Pulse transit Method to improve the sensitivity,response and reduction of losses has been worked out. Thispaper discusses the signal processing technique and electronicsdesign approach for development of this Snow Depth Sensor.I. INTRODUCTIONThe importance of snow hydrological studies andmeasurements has been recognised globally and anintegrated approach to study of all the interdependentparameters is underway and related sensors are beingdeveloped [1]. The snow cover thickness is one of the mostimportant hydrological parameters, which is used in theforecasting model for snow avalanche release, river runoff water, glacier sliding and related phenomenon in themountain areas and planes nearby[2, 22]. In the deep snowbound areas, the snows cover thickness changescontinuously because of the radiation, heat and climaticchanges. The snow cover thickness is an indication of snowbound water.A Snow Depth Sensor system has been designed [3]tomeasure the snow depth in highly snow bound areas. Thissystem records the snow depth after suitable interval. Thisdata is integrated with the automatic weather station to studythe variations in the snow cover thickness, net melting andrun-off water.Pulse transit method employed in snow depth sensorsystem: The thickness of the snow cover is found with thehelp of pulse transit method[4]. A short burst of ultrasonicpulses is transmitted by the piezoelectric transducertransmitter, which is mounted on a pole with the sensorsfacing vertically downward towards the snow. Refer to Fig[1]. The transmitted beam strikes the surface of the snow.Some of the energy get reflected back and is received by thereceiver. The time of travel between the transmission andthe reception of the pulses is computed which gives thedistance of snow cover from the sensor as per distance-velocity equation. The mounting height (distance of sensorfrom the ground) is already known. By subtracting themeasured distance from the installation height, snow depthis obtained.Snow has critically irregular and non-smooth poroussurface, which cause penetration of wave of incident waveinto the surface, scattering around and absorption in thesurface results in very low amplitude highly noise riddenreflected signal[21]. This received echo signal has to beconditioned to remove signal and processed to increase itsamplitude to make it sufficiently detectable. Sometime thereflected echo is not received back at all and design has tobe carried out to increase the probability of receiving backthe reflected echo. A special design has been worked out tocompensate for the losses due to snow.Fig [1]: Snow depth sensor mounting ArrangementThis paper discusses the design of Snow Depth Sensor toimprove the signal processing and sensitivity as well asrange of the system. Further design for near fieldcompensation as well as compensation for variation ofsound velocity with temperature is also highlighted.II. DESIGN CONSTRAINTSThis simple pulse-transit method of finding the thickness isnot as simple when applied to the snow. There are manyfactors like losses in snow, temperature at the place of themeasurement and irregularities of snow surface, whichgovern the range and reliability of the snow depth sensor.These have to be taken into consideration while designingthe system.Extreme Environmental constraints: One of the biggest
  • 2. R. Attri Instrumentation Design Series (Snow Hydrology), Paper No. 3, August 1999Copyright © 1999 Raman K. Attri Page 2problems in the system design is the temperature andenvironment specification. Temperature encountered at theplace of installation is of the order of -40oC and relativehumidity of 100% in heavily snow bound area. Dependenceof ultrasonic velocity on temperature and change indirectional pattern of the sensor due to fluctuation intemperature is also a big problem.Low Amplitude of reflected echo: The overall performanceand reliability of the snow depth sensor based on thismethod depends upon the ability of the system to detect thereflected echo of the transmitted signal. Further the systemperformance can be increased only if one can enhance thereflected echo to a sufficiently detectable level.Irregularity and non-smoothness of snow surface: In thisapplication, the shape and the roughness of the surface isof decisive importance. These factors often limit thesensitivity of snow depth sensor. A roughness of more than1/10 of wavelength impairs the coupling markedly. Therough surfaces make the ultrasonic beam to become diffusedand scatter in all directions.Porosity of snow cove: The porosity of the snow causeslarge amount of ultrasonic energy to get absorbed in thesnow, so only a small part of it get reflected back. Thereceived echo is very weak which need considerable amountof amplification. The worst of all is that the ultrasonic beamis striking the highly non-smooth surface, which causes thescatter, and the missing of reflected beam. As a resultsystem has to be very sensitive to detect the weak reflectedecho as well. The reflected echo is around 0.1 mV inresponse to transmitted wave of 100 V.Noise ridden signal: The low amplitude reflected echo isfurther over-ridden in white noise. As the ultrasonic wave ismechanical wave, so signal noise is very much mechanicalin nature spread over the entire spectrum. Noise amplitudeis more than the signal amplitude so system with largersignal-to-noise ratio has to be designed. Inherent residualnoise of the piezoelectric Transducers itself is great sourceof problem.The worst of all is that sometime the reflected echo is notreceived at all. The reflected beam reflects at such angle thatit completely misses the receiver sensors and systemremains in wait mode.Experimentally it has been found that increase intransmitted voltage amplitude do have some improvementof results, but this improvement get saturated after certainpoint and no amount of increase in transmitted voltage makeany effect. Further the increase in amplification of receivedecho will only increase the noise beyond the threshold level.The use of lower frequency does have some positive effectsof increasing the range and reducing the losses due to scatterand absorption, but there is a limit on lower frequencyultrasonic transducers available.So a hardware-software co-design approach has beenworked out to modify [1]the directional pattern of thetransmitting beam in such a way to reduce the scatteringeffect and attenuation of the reflected beam. This designfurther increases the sensitivity of the system by signalprocessing to extract the useful reflected echo and detectionof missed reflected echo.Inherent errors Associated with ultrasonic beamAlong with design constraints, some errors are alsoencountered in the system. It pertains to the near-fieldinterference and grating lobes caused by this interference [5].The reflected echo is received as result of these gratinglobes instead of the actual reflection, thereby causing effectof side-looking, implying as if target is lying in theimmediate proximity of the transducers [6]. This near-fieldcompensation has to be provided. The other error is causedby variation of ultrasonic velocity with the temperature.This velocity compensation has to be provided in the designof the system.III. OVERVIEW SNOW DEPTH SENSORSYSTEM ARCHITECHTUREThe block diagram of the system is given in the fig[2]. Thesnow depth sensor have three section namely I) TransmitterSection ii) Receiver Section iii) Control SectionFig [2]: Detailed system block diagram of Snow Depth SensorThe system is auto power on triggered and hence does notneed any switches and buttons to control the system. Thiswill give the measurements as soon as the power is suppliedto the system and every 5-Sec after the power on. Thepower is usually provided by the Data Acquisition systemdepending upon its requirement of measurement and itssampling interval. As soon as the reading is taken, power isswitched off. The circuit operates under the total control ofsoftware. All the pulses are generated by software inconjunction with the hardware circuitry for triggering of themeasurements through peripheral devices. It contains therequired circuitry needed for ultrasonic transmission andsignal conditioning of received ultrasonic signal. This
  • 3. R. Attri Instrumentation Design Series (Snow Hydrology), Paper No. 3, August 1999Copyright © 1999 Raman K. Attri Page 3includes input impedance matching circuit, high gainamplifiers and narrow band pass filters. Further the receivedpulses are shaped and converted into suitable format. Thenear field compensation circuit is also incorporated here.Temperature transducer AD590 has been used in this boardto sense the temperature and provide corresponding voltageand ADC is used for converting this voltage into digitalword and provides compensation for the sound velocity.This board also contains the DAC to provide the conversionof computed results into real word analog and frequencysignals for interfacing with any kind of Data-loggers. 12 bitADC and 12 bit DAC is incorporated to get sufficientaccuracy of measurementsThe dual output is taken from the system. One is analogvoltage in the range of 0-1 V and other is frequency in therange of 0-10 kHz in proportion to the distance in the range0- 10 meters. This output is interfaced to data collectionplatform or a data acquisition system for further processing[20].Provision for testing modules is also provided and to takemanual measurements. Raw data is further computed usingextensive software techniques to convert it into usable dataoutputThe power supply card of this sensor also take care ofgenerating high voltage needed for operation of thepiezoelectric transducers, in addition for providingconventional power supply for the systemIV. DESIGN APPROACHThe design efforts have been made towards improving:i) extraction of 40 kHz signal from white noiseii) dynamic rangeiii) signal processingiv) wave-shaping and pulse-shapingv) preciseness and repeatability of distancemeasurementsvi) range of snow depth measurementvii) functionality in low temperature rangeviii) Universal standards of output formatsThe design approach consists of modifying the directionalresponse of the sensors, getting the effective transmission ofultrasonic energy and increasing the sensitivity of thereceiver. This design approach is achieved by proper signalprocessing, noise removal and signal conditioning,amplification and filtering backed by proper software-hardware co-design.Design for improved directional response: The first mostthing is to choose proper ultrasonic generating transducers.For present application piezoelectric transducers operatingat 40 kHz have been chosen. The higher frequency causespenetration in the porous snow surface and cause lot ofheating and absorption in the surface[7]. The range ofmeasurement is very poor at higher frequencies. Frequencylower than 40 kHz is quite suitable as far as directionalpattern and other losses are concerned, but then the accuracyof the measurement suffers. Further the availability of lowerfrequency moisture proof completely sealed piezoelectriccrystals working in –50 to +50 degree temperature range isquite difficult. The experiments have been done to select theproper sensors. Refer to fig [3] which compares the range,accuracy and penetration of ultrasonic beam frompiezoelectric sensors at different available frequencies [8].Penetration is maximum at 50 KHz and accuracy is bestwith 25 KHz. Taking into consideration the range, accuracyand penetration inside the snow. 40 KHz crystal gives theoptimum performance, which has been selected.Fig [3]: Graph of values of range, penetration and relativeaccuracy of Piezo-electric crystals as function of frequencyThe Hall effect produces longitudinal mechanical wavesalong the axis of the crystals. The specifications of thePiezoelectric Crystal used are given in the table [1].Table[1]: Specification of the PZT ultrasonic transducersParameter ValueModel MA40E6-7 or MA40E7S-1Nominal Frequency 40.0 khzSound pressure level > 108 dB at 40 khz at 30 cm , 10Vrms(Sine wave)Sensitivity > -82 dB at 40 khzCapacitance 2200 pF+ 20% at 1khzFurther the mounting arrangement of these sensors is notconventional. Although the crystals used are reversible (canbe used in transceiver mode), even then separate sensors fortransmitter as well as receiver are used. Further one-to-oneratio of transmitting and receiving sensor has not been foundsuitable in the present application.In order to deal with the non-smooth surface of snow and tocounteract the scattering effects, the directional pattern ofthe transmitted ultrasonic beam has been modified using
  • 4. R. Attri Instrumentation Design Series (Snow Hydrology), Paper No. 3, August 1999Copyright © 1999 Raman K. Attri Page 4Ultrasonic Transducer array arrangement. The beam patternhas been made relatively broader with large front zone area[9]. This design effort increases the possibility of gettingeven the scattered beam from the non-smooth surface ofsnow. The configuration consists of separate 3 x 2 receiverarray and 3 x 1 transmitter array isolated from each othermechanically as well as electrically. The configurationimproves the directional response and beam widthconsiderably and optimum performance can be achieved byvarying the geometrical parameters as well as frequency oftransmission. Refer to fig [4] which shows the directionalresponse of the Array Transducers.[10]Fig [3]: Directivity of MA40E7S-1 transmitting piezo-electrictransducer in terms of sound pressure level measured in dB at30cm with 40 KHz frequency. Typical directivity is 75osymmetricin all directions. This results in a broader beam with broad frontzone area of the beam (Source: Murata Corp Japan) [19]Design for improved transmission: The primary responseof the snow depth sensor depends upon the amplitude of thetransmitted signal. Greater the amplitude, better shall be thereflected echo. Normally piezoelectric ultrasonictransducers are operated at high voltages as high as 100 V.Experimentally it has been observed that increase intransmitted amplitude have very minimal effect on theresponse of the system. With increase in 10% of transmitteramplitude voltage only 1% improvement in the response isachieved. So transmitter voltage can not be raised beyond athreshold, if raised at all, the improvement of response isminimal at the cost of heavy current consumption. 60 V hasbeen chosen as optimal transmitter voltage after manyexperiments. The circuit diagram of the transmitter sectionis as shown in the figure [5].A single pulse triggers the transmission section. Transmitteroperates on 40 kHz frequency generated by 40 kHzoscillator. In pulse-transit method, only a short burst ofpulses is sent toward the target. The number of pulseschosen for transmission can be critical depending upon theapplication [11]. Since target here is quite irregular surfacewith discontinuities, number of pulses has to be more thanthe generally used. Experiments have been done to find outoptimal number of pulses for transmission. At the receiverside the transmitted pulses make a modulated kind ofenvelope with peak at centre of the envelope. The pulses atthe extremes of this have very low amplitude. 16-pulsepattern has been found to the most suitable pattern as a burstof ultrasonic.Fig [5]: Transmitter Section is a Class-C amplifier operating at60V, driven by a switching transistor. An oscillator and RS flip-flop combination logic ensures passage of 16 pulses to poweramplifier and to series array of transducers.A R-S flip-flop in conjunction with the 16 pulse counter isused to transmit only 16 pulses to the base of TransmitterAmplifier. The transistors are being operated in switchingmode. The output stage is class-C power amplifier operatingat +60 V. The transmitting transducers are coupled to outputstage through a coupling capacitor. Precaution is taken notto apply direct DC voltage to the piezoelectric transducersfor a long time. The duty cycle of 25 % is chosen to givetolerance for the transition time of transistors and rise timeof ultrasonic transducers.Since the longitudinal mechanical waves generated from thepiezoelectric sensor travels in the medium towards the targetand received echo is reflected back and travels towards thesensor. Refer to Figure [6a], which shows what is beingtransmitted. The transmission envelop is shown only for onePZT crystal, while in actuality there are three envelopssuperimposing each other. This is more clear in Fig [6b]showing received echo waveform envelop.Design for improved sensitivity at receiver section: Thereceiver section has been designed with extreme carebecause the whole performance of the system depends uponthe fact that how accurately the reflected echo is detected bythe system [12]. The reflected signal received at the output ofthe piezoelectric sensor is as minute as 1 mV in response tothe 60V transmitted signal. This reflected echo is shadowedso heavily by the noise of higher amplitude that it isimpossible to detect it without signal conditioning. Thereflected echo is applied to the input amplifier stage whichserver the dual purpose of impedance matching as well asinitial gain. Refer to Fig [7] showing the receiver circuitdiagram in detail. Transistor self-bias with quotient point inthe middle of the load line is used as input impedancematching stage. Signal is coupled to this stage through acoupling capacitor, as the signal of interest is alternating at40 kHz. Q-point selection is extremely important here, as nopart of the signal should get trimmed and faithful amplified
  • 5. R. Attri Instrumentation Design Series (Snow Hydrology), Paper No. 3, August 1999Copyright © 1999 Raman K. Attri Page 5signal should be provided by this stage.Fig [6a]: A burst of 16 ideal square pulses of 60V triggered byoutput amplifier results in modulated mechanical waves at thetransmitter transducer. The received signal also exhibit similarmodulated characteristics.Fig [6b]: 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 sequenceGain provided by input stage is 200. The amplified signal isstill highly noise ridden, noise spanning over the entirespectrum [13]. The output signal as shown in figure [8a] isstill in mili-volts range. A high gain amplifier is used forenough amplifications of the signal. An optimum gain ofapproximately 50 has been given at this stage. It is usuallynot possible to give very high gain at this stag because ofGain-Bandwidth product constraints. The useful signalbandwidth is approximately 50 kHz, which implies unitygain bandwidth of approximately 2500 kHz. A videoamplifier with UGB of 120 MHz has been used to amplifyall the frequencies uniformly. The output signal now is ofthe order of 1 V, but still shadowed by the noise of higheramplitude. Experiments show that the high amplitude noiseis well above or well below centre frequency of 40 kHz. Asshown in fig [8b] the reflected echo is still not reproducedfaithfully.Fig [7]: Block diagram of the receiver wave shaping circuit. Thecircuit amplifies the weak received signal, compare it with anominal threshold to eliminate the white noise and shape theresultant signal into square wave pulsesA precision filter amplifier has been designed as narrowband pass filter with sharp cut-offs at 38 kHz and 42 kHzwith centre frequency of 40 kHz. The detection of thereflected echo and hence the sensitivity of the measurementis very much dependent upon the accuracy of centrefrequency of this filter. This is a kind of notch band passfilter, which should reproduce only 40 kHz frequencyfaithfully [14]. All other frequencies are filtered out. Theresponse of this filter stage is shown in the figure [8c]. Theresultant echo is of very clear shape and amplitude. Thenumber of pulses received in this echo is three times thenumber of transmitted pulses (because of three transmitters).Actually the received echo consists of three envelopesshifted slightly in phase and each envelope having 16pulses. The increased number of pulses gives better chancesof triggering of next circuit and hence the measurement andhence better sensitivity.The pulses received are conditioned and wave-shaped todrive the digital circuit, refer to fig [8e]. First pulse out ofincoming pulse stream pattern (Fig [8f]) is extracted as onlythis pulse corresponds to arrival of the echo itself. Firstpulse extracted from transmitted stream has started a timingcounter and this first pulse of incoming stream stops thiscounter.The counts in the counter are direct indication of flight time(time taken by the echo to travel to target and return back tothe sensor). This time is used in computation of distance oftarget surface from the sensor as described in the beginningof this paper.
  • 6. R. Attri Instrumentation Design Series (Snow Hydrology), Paper No. 3, August 1999Copyright © 1999 Raman K. Attri Page 6Fig [8]: 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 reflectedecho. First saturated signal corresponds to either side lobes or false trigger due to broader beam c) Output signal at Filter Amplifier outputwhich cleans the signal and removes the noise. The reflected echo signal is enhanced for digitization. d) Pulse stream at comparator whichremoves the background residual noise and digitizes the inputs signal peaks above the threshold voltage e) Pulse stream at Wave shaperOutput removes the negative side of the pulses and produces clean rectangular pulses from 0-5V f) Output of 16-to-1 pulse isolator circuitwhich produces one pulse against a stream of 16 pulses. This single pulse is used to stop the timing counter for measuring transit time
  • 7. R. Attri Instrumentation Design Series (Snow Hydrology), Paper No. 3, August 1999Copyright © 1999 Raman K. Attri Page 7One important 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. 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 crossconvolution techniques are employed to remove suchresidual noise. In present case alternative method ofcomparison with some fixed threshold has been used. Theaverage value of residual noise is checked out. Acomparator circuit as shown in figure [7] is used to comparethe filtered output to this threshold voltage [15]. If thereflected echo above this threshold, comparator gives theoutput in the form of pulses crossing the threshold mark.The comparator also wave-shape the bipolar sine signal intouni-polar square pulses.One important precaution in this design is the setting of gainand threshold value of the circuit. A trade-off point havebeen observed beyond which accuracy may get impaired ifwe further lower the threshold and increase theamplification in order to increase the system sensitivity.Once we have obtained the pulse corresponding to start oftransmission and that corresponding to reception, we cancalculate the flight time either using pure hardware or usingsoftware. Both the methods have been implemented.Software in assembly language is a better choice as it givesadded advantage of calculation and system control together.The half the flight-time when multiplied to velocity ofsound at that temperature gives the distance of the reflectingsurface from the sensor. By subtracting this distance fromthe mounting height we get the snow depth at that time.V. SYSTEM PERFORMANCE EVALUATIONThe system has been designed to work not only forsampling based measurements, but also for continuousround the clock functioning as well. The system has beenimplemented in single mechanical unit for easy handlingand installation. The mechanical chassis and design isabsolutely very important aspect of the overall engineeringof the system. The system is expected to work round theclock in the heavily snow bound areas and continuouslyinvaded by the harshest possible environment. The systemstill has to give the accurate reading of the snow depth inthis type of hostile environment as well. The componentfailure and system resistance to moisture may adverselyaffect the performance of the system. As the system are tobe installed at those places where frequent man visits andrepairs etc are not possible. So system has to undergo manytough testing. The three performance tests are used over thissystemi) long term stability testingii) environmental testingiii) integrity testingThe system is sealed into a waterproof chassis and put intooperation in a cold chamber at –40 degree centigrade for 24hours. These tests are in conformation with the J55555military specifications.The long-term stability of the readings is checked byrecording the distance shown by this system continuouslyfor 24 hours. The fix target distance shown by this sensor ischecked to be same during these 24 hours.The total system when interfaced with the automaticweather station or data collection platform is tested foraccurate distance at all level of the moving target[16]. Thecable drop through which it is interfaced to the data-loggeris chosen such a way that it gives minimum voltage dropover the length of the cable.The overall performance and accuracy of the system alsodepends upon the components and the PCBs. Militaryspecification JM38510/ JM883-B components have beenused which can operate in the temperature range of –60degree centigrade to +125 degree centigrade. Gold platedPCB has been designed. The military standard weatherproofconnectors and cables have been used.The software in conjunction with the hardware is used toimprove the system performance. The accuracy of + 1 cmwith the resolution of 1 cm is achieved for the range of the 5meters. Snow depth upto 5 m above the ground can bemeasured accurately with this system. System performanceover all kind of porous and irregular surfaces has beenchecked out and has been found very reliable.Fig [9]: 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 itmoves away from 450cm, the scattering effects comes into picturewhich may make the reflected beam to miss the receiver altogether.The accuracy is impaired in far field. Detection of surface isinhibited up to 50cm in the near field due to side lobecompensation (the captions on the points indicate the measuredabsolute distances in cm)
  • 8. R. Attri Instrumentation Design Series (Snow Hydrology), Paper No. 3, August 1999Copyright © 1999 Raman K. Attri Page 8System performance diagram in Fig [9] shows the error incm at different points in the range of the sensor. The wave issometime missed and not received back as echo if the senoris away from target surface more than 450 cm (far field). Itdoes give the indication of out of range target surfacethrough software programming. On the other hand near fielderror comes in picture due to side lobe compensation andsystem do not detect any surface when it is in proximity of50 cm to the sensor face. The side lobe compensation circuitsacrifice near field 50 cm measurement range, whichpractically is never used. The snow surface hardly reacheswithin 50 cm proximity of the sensor. The system accuracyis within + 1 cm range.VI. DESIGN USED FOR COMPENSATION OFOBSERVED ERRORS AND IMPROVEMENTOF SYSTEM RELIABITYThe major factor, which affects the system performance, hasbeen the variations of the sound velocity with temperature.This compensation has to be provided. An integrated solidstate temperature sensor AD590 has been used along-withits linear circuit, which reads the temperature with theresolution of 0.01 degree centigrade. 12 bit ADC is used fordigital conversion of temperature. Extensive Softwarecomputations have been carried out to compensate theeffects of the temperature over the velocity of the sound.While checking the system performance, it has beenobserved that the two streams of the received pulses isobtained, first one with large amplitude. The first streamcomes immediately after the transmission and other streamarrives late. The first stream of pulses is corresponding tothe side-lobes of the transmitted beams, produced as a resultof the near-field interference[17]. This fist pulse will producea counter stop pulse, which shall stop the timing counterprematurely. This effect is called side looking and systempresume as if an object is lying in the immediate proximityof the sensor. First stream of pulses is blocked by hardwareas well as software implementation. The figure [10] showsthe circuit for blocking the first stream of pulses till the totalwidth of first stream so that the sensor responds to actualecho only.Fig [10]: Near-field compensation circuit to stop the systemcounter to avoid false triggering of receiver circuit due to broadbeam patter from transmitter and / or side grating lobes.The voltage level at input of OR gate is made high till theexpected time of first steam. The expected time has beenaround 2300 microseconds. A safe margin is given andpulses till the time 2560 microsecond are blocked. After thatany stop pulse arriving shall be passed through and the stopthe counter. In software any reading below countscorresponding to this time are rejected and measurement istermed false and cycle is again repeated. This is near-fieldcompensation.In the beginning of the discussion it has been highlightedthat if the snow surface is much beyond the range of thesensor, then the reflected signal echo will be much belowthreshold level and it will not be detected. So system timingcount shall keep on increasing without any intelligentsensing of missed echo. The situation shall be same becauseof the severe scattering echo beam arrive the sensor at suchan angle that it completely miss the sensor. In this casesoftware checks for the maximum expected counts. Ifcurrent counts crosses the maximum permissible counts,system take the situation as missed beam and repeat themeasurement cycle until it gets the counts in the permissibleaccurate range. Even when surface is beyond its range,during some measurement cycle it has been found thatincident beam strikes the surface at such an angle that thescattered beams make high amplitude interference pattern atthe receiver sensor. This give rise to detectable echoamplitude even when surface was beyond range of thesystem. This far-field compensation enhances the systemperformance and makes it intelligent enough to detect thepossibility of surface in the path..VII. CONCLUSIONSThe experimental results have been very favorable.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 has been enhanced usingthese design techniques. The snow surface distance morethan 5 meters can be fairly detected inspite of the non-even,rough, porous snow surface which normally do givesproblems of scattering and absorption of energy in to thesnow. The directional patterns have been made broader tooffset the effects of non-smoothness of the surface. Thereliability of snow depth parameter, which is measured bythis snow depth sensor, is extremely important for modelingof snow avalanche forecast and other related studies [18, 19].This snow depth data is recorded after suitable intervals bythe data collection plate form and keep the track ofdeposited snow and snow melt.ACKNOWLEDGEMENTSDr. B.K. Sharma, Head Of Dept, Geo-Scientific InstrumentsDivision, CSIO, ChandigarhSwaranjit Singh, Senior Technical Officer, Geo-ScientificInstruments Division, CSIO, Chandigarh
  • 9. R. Attri Instrumentation Design Series (Snow Hydrology), Paper No. 3, August 1999Copyright © 1999 Raman K. Attri Page 9REFRENCES1. Shamshi, M.A., Attri, R.K., Sharma,V.P., Snow PackTemperature sensor, Proceedings of National Conference onSensors and Transducers, pg. 180-189, 1996.2. Mellor, M., Engineering properties of snow, Journal ofGlaciology, volume 19, pg 15-66,19773. Satish Kumar et al, Snow Depth Senor, Proceeding of nationalsymposium on Sensors and Transducers, 1996.4. Gooberman G.L., Pulse Techniques, Ultrasonic Techniques inBiology and medicine, Illiffe books Ltd, London, 1967 pp 87-128.5. Balantine D.S et al, Acoustic Wave Sensors-theory, design andphysico-chemical applications, Academic Press, 19976. Busch I, Huxson E, Encyclopaedia of applied Physics, VCHPublishers, vol 1, pp 63-88, 19917. Ensminger D, Ultrasonic – the low and high intensityapplications , Marcel Decker Inc, New York, 19738. Krautkramer L and Krautkramer H., Ultrasonic Testing ofmaterials, Berline: Springer, 1969.9. Encyclopedia of Science and Technology, Vol 12, pp 662-664,Mcgraw Hill, 1982.10. Hudson J.E, Adaptive Array Principles, Peter Peregrinus ,London 198111. Silk M.G., Ultrasonic transducers for non-destructive testing ,Adam hilger Ltd, Bristol, 198412. Zhang D, Crean G.M., Non-linear acoustic properties of highpurity quartz as function of temperature, Proceedings ofDevelop,ment in Acoustic and Ultrasonic, IOP PhysicalAcoustic Group , Leeds, 24-25 Sept 1991, pp 219-22413. Landee, R.W. et al, Electronics Designer’s handbook, McgrawHill , 195714. Raman S et al, Processing of Ultrasonic signal for transducercharacterisation and for improving signal to noise ratio, IndianJournal of Technology, Vol 31, Nov 1993, pp 774-77615. Lawson, J.L. and Uhlenbeck, G.E., Threshold Signals, McgrawHill, 1948 pp 21116. Ganju A, Snow cover model, Proceedings of SNOWSYPM-94,pg221-226,sept 199417. Szilard J, Ultrasonic testing – non-conventional testingtechniques, Willey, NY, 198218. Jordan, R.. A one dimesional model for snow cover, CRREL,Special report 199119. Yamazaki, Kondo, T. J., Sakuraoka,T. and Nakamura, T., Aone dimensional model of evolution of snow covercharacteristics, Annual of Glaciology, 18, pg 22-26,1993,20. Attri Raman K., Sharma B.K., Shamshi M.A., Practical DesignConsiderations for Signal Conditioning Unit Interfaced withmulti-point Snow Temperature Recording System. IETETechnical Review, Vol 17, No 9, pp 351-61 Nov-Dec 200021. Papadakis E.P., Physical acoustic principles and methods,(W.P. Mason , ed.), vol 4 B, Academic, NY, 196822. Hall, Derothy K., Remote sensing of Ice and Snow, London,Chapman and Hall Publications, 1985,Aut hor Det a ils :Author is Global Learning and Training Consultantspecializing in the area of performance technology.His research and technical experience spans over 16years of project management, product developmentand scientific research at leading MNC corporations.He holds MBA in Operations Management, ExecutiveMBA, Master degree in Technology and Bachelordegree in Technology with specialization inElectronics and Communication Engineering. He hasearned numerous international certification awards -Certified Management Consultant (MSI USA/ MRAUSA), Certified Six Sigma Black Belt (ER USA),Certified Quality Director (ACI USA), CertifiedEngineering Manager (SME USA), Certified Project Director (IAPPM USA), toname a few. In addition to this, he has 60+ educational qualifications,credentials and certifications in his name. His interests are in scientific productdevelopment, technical training, management consulting and performancetechnology.E-mail: rkattri@rediffmail.comWebsite: http://sites.google.com/site/ramankumarattriLinkedIn: http://www.linkedin.com/in/rkattri/Copyright InformationCopyrights © 1999 Raman K. Attri. Paper can be cited withappropriate references and credits to author. Copying andreproduction without permission is not allowed.