31-12-2009, 05:56 PM
[attachment=922][attachment=923][attachment=924]
ABSTRACT
Machine health monitoring is of critical importance to its safe and reliable operation. The emergence of wireless data transmission techniques has extended the Scope of health monitoring to machine components and systems that are difficult to accessor not suited for wired sensor data acquisition .this paper presents the design and prototype realization of a digital wireless data transmitter based on the CDMA technique.
The design provides a generic solution for various applications in electronic instrumentation measurement and embedded sensing for machine condition monitoring and health diagnosis.
1. INTRODUCTION
The increasing demand on productivity and quality has led widespread automation of manufacturing processes and equipment. Cost effective and efficient maintenance is needed to prevent costly machine failure related downtime and ensure product quality. For the past decade the state of technology for machine health monitoring has been continuously improving. But majority of monitoring systems employed at the factory floor still require maintenance engineers to manually collect data and analyse them off-line. Most of these systems require wire connections for data acquisition and transmission ,which limit their accessibility and utility in many situations where space restrictions do not allow for such connections.
Wireless data transmission for embedded sensing and machine condition monitoring has been developed in recent years to overcome the restrictions of wired data connections. Because of the constraints concerning power supply, most of these systems operate on batteries and transmit data over a short distance to a data logging station nearby. A challenging issue in wireless data transmission is to design for low power, less circuitry complexity, and high reliability.
A low power ,compact digital wireless data transmitter employing code division multiple access scheme has been introduced designed simulated, prototyped and bench tested in lab environment. The design provides a generic wireless solution for embedded sensing and measurement. Salient features of the transmitter include an interface to multiple sensors ,and sharing of the save data receiver by multiple transmitters. Using surface mount packaging ,all the components of the transmitter could be accommodated within an area of about half the size of credit card,making transmitter well suited for structural integration into machines. Since the components are all available in the form of ICdies, the entire design can be further miniaturized into a single hybrid chip.
2. SIGNAL MODULATION TECHNIQUES
Inorder to transmit data wirelessly in an air channel ,amodulation scheme is needed inorder to translate data symbols to variations of acarrier wave of specified transmission frequency .Three major techniques exists for digital data modulation are: ASK, FSK, and PSK .The ASK technique uses amplitude variations of the carrier wave to represent the symbols transmitted, FSK uses different frequencies for each symbol, and PSK uses the phase shift of the carrier for symbol representation.
The robustness of wireless data transmission is measured by the probability of a symbol error, which is dependent on SNR that measures the strength of transmitted signal with respect to that of background noise.
A. Amplitude Shift Keying
The ASK technique modulates the data by assigning each symbol a different amplitude level, eg: two levels for binary data. For application in a metallically sealed machine environment such as a bearing housing where signal reflections may superimpose each other, an ASK receiver may not be able to distinguish between the original signal and the reflected one ,leading to data misinterpretation. Hence the ASK was not considered for the present application.
B. Frequency Shift Keying
This technique employs different frequencies for different symbols transmitted. To ensure reliable transmission, the difference between two frequencies used has to be at least
where T is the duration of the symbol.
FSK is a low performance type of digital modulation. Binary FSK is a form of constant amplitude angle modulation. The general expression for binary FSK is
V fsk(t) = Vc COS { 2 [ fc+ Vm(t) ft}
Where V fsk(t) = binary FSK waveform
Vc = peak carrier amplitude ( volts )
fc = carrier centre frequency ( hertz )
f = peak frequency deviation ( hertz )
Vm(t) = binary input modulating signal with binary FSK, the carrier frequency is shifted by binary input signal.
For binary data transmission, the probability of error is
where represents SNR PER BIT.
Employing symbols that comprise more bits will reduce error probability, but will increase complexity in filtering and decoding hardware on receiver end. Such complexity would further lead to increased power consumption, which is detrimental to the bearing condition monitoring applications so the FSK technique was not considered for the present design.
C. PHASE SHIFT KEYING
Here , the carrier wave by itself represent a symbol, whereas all other symbols are defined by phase shift from the phase of carrier.
Phase shift keying (PSK) is a form of angle-modulated, constant-amplitude digital modulation.
With binary phase shift keying (BPSK), two output phases are possible for a single carrier frequency (binary meaning 2). One output phase represents a logic I and the other a logic 0. As the input digital signal changes state, the phase of the output carrier sifts between two angles that are 1800 out of phase.
For a binary signal the phase shift for second symbol will be180 degree. In fig 1, a binary data signal 110101 is illustrated, together with the carrier wave and the modulated signal. The logic level 0 is represented by -1, since the receiver cannot differentiate no transmission and transmission of 0. whenever the logic level of data changes, a phase shift of 180 degree occurs in the modulated signal. So in PSK the carrier wave by itself represents a sybol and all other symbols are defined by the phase shift of carrier phase
Fig. 1. Datatransmission of a string 110.101using PSK modulation
The probability of a symbol error for transmitting a binary data using PSK is
If the quadrature phase shift key is used that transmit two bits at the same time, then the error probability is
For an assumed SNR of 8 dB, the probability of error transmitting a symbol using QPSK is 0.02% and for PSK is 0.01%.To compare them on a bit error basis the symbol error probability has to be translated into bit error probability. The relationship between symbol and bit error probability can be approximated as
Pebit =1/2 Pe
This means that with QPSK it is possible to transmit twice as much data within the same period of time as with ordinary BPSK with same low probability of error. This result cannot be transferred simply to other modulation variations that combine more than two bits in one symbol.
Quaternary phase shift keying (QPSK), or quadrature PSK as it is sometimes called, is an other form of angle-moduled, constant-amplitude digital modulation. With QPSK four output phases are possible for a single carrier frequency. Because there are four different output phases, there must be four different input conditions. Because the digital input to a QPSK modulator is a binary (base2) signal, to produce four different input conditions, it takes more than a single input bit. With two bits, there are four there are four possible conditions: 00, 01, 10 and 11. There fore, with QPSK, the binary input data are combined into groups of two bits called dibits. . Therefore, the rate of change at the output (baud rate) is one-half of the input bit rate.
QPSK transmitter
A block diagram of a QPSK modulator is shown in figure 2. Two bits (a dibit) are clocked in to the bit splitter. After both bits have been serially inputted, they are simultaneously parallel outputted. One bit is directed to the I channel and the other to the Q channel. The I bit modulates a carrier that is in phase with the reference oscillator (hence, the same I for in phase channel), and the Q bit modulates a carrier that is 900 out of phase or in quadrature with the reference carrier (hence, the name Q for quadrature channel).
It can be seen that once a dibit has been split in to the I and Q channels, the operation is the same as in a BPSK modulator. Essentially a QPSK modulator is two BPSK modulators combined in parallel. Again, for a logic 1 = + 1V and a logic ) = - 1 V, two phase are possible at the output of the I balanced modulator and two phases are possible at the output of the Q balanced modulator When the linear summer combines the two quadrature (900 out of phase) signals, there are four possible resultant phasors given by these expressions:
With QPSK each of the four possible output phasors has exactly the same amplitude. Therefore, the binary information must be encoded entirely in the phase of the output signal. This constant amplitude characteristic is the most important characteristic of PSK that distinguishes it from QAM. Also, from figure 2 b it can be seen that the angular separation between any two adjacent phasors in QPSK is 900 . therefore, a QPSK signal can undergo almost a+450 or a- 450 shift in phase during transmission and still retain the correct encoded information when demodulated at the receiver. Figure 12-21 shows the output phase-versus-time relationship for QPSK modulator.
Each dibit code generates one of the four possible output phases. Therefore, for each two-bit dibit clocked into the modulator a single output change occurs.
Employing QPSK instead of BPSK requires additional operation steps, either by adding specialized h/w components or a software implementation .
3. MULTIPLE ACCESS METHODS
In order to accommodate multiple transmitters with one data receiver and achieve design flexibility and extensibility, a multiple data access scheme needs to be implemented. This can be achieved by assigning each transmitter a distinct feature eg.,certain frequency a time slot or a code .Accordingly , three design options were considered in the present design :frequency division multiple access[FDMA],time division multiple access [TDMA]and code division multiple access [CDMA] techniques
A. FDMA technique
In FDMA,each transmitter has a frequency exclusively assigned to it. This enables transmitter to send data at any time regardless of other transmitters. The frequency ranges can be transmitted simultaneously. The frequencies should be non-overlapping. If overlapping one of the overlapping range is shifted to another range with same bandwidth.S ince absolute real time monitoring systems are usually not needed in production environment,a transmitter will need to send data only at specified intervals,leaving the assigned frequency unused most of the time.
B. TDMA technique
A single carrier frequency can be used by several transmitters, if it is ensured that at no time more than one transmitter sends data simultaneously . Accordingly, each transmitter is assigned a time slot for signal transmission .The time slots need to be synchronized between all the transmitters in the system. This requires that for each transmitter, a separate receiver is implemented, and the receiver must be ON at all times, increasing the power consumption of the circuitry.
C. CDMA technique
It is a digital cellular technology. It uses spread spectrum technique. CDMA doesnâ„¢t design a specific frequency to each user. Each channel use fully available spectrum. Individual conversions encoded with pseudo random digital sequence. It is a military technology. It was used in second world war in jamming transmission. Using this all stations are permitted to transmit over the entire frequency all the time. Multiple transmissions are separated using coding which makes an assumption that when multiple signals combined together they will not get gargled. They add linearly.
Each bit is sub-divided into ˜ m™ short intervals called chips. Each station is assigned to a unique m “ bit code called chip sequence. To transmit a 1-bit the station will be transmitting the chip sequence and for a o-bit the station will be transmitting 1™s compliment of chip sequence. No other patterns are permitted. We use bipolar signals with 1-bit being +1 and 0-bit being -1. Suppose a station is having a chip sequences. An important assumption to be noted is that all chip sequences must be pair wise orthogonal is the normalized inner product of any two different chip sequences must be zero ie, S.T=0 and S.T=0. where T another stations chip sequence.
Recovery of data: to recover the data being transmitted the receiver must know, the chip sequence of the sender in advance. So for recovery process, the normalized inner product of transmitter and receiver is found. If S is the received chip sequence and C is the chip sequence of sender, three cases are.
i. S.C = 0 ( C has not transmitted anything )
ii. S.C = 1 ( C has transmitted a 1-bit )
iii. S.C = -1 ( C has transmitted a 0-bit )
This has been divided into frequency “ hopping [FH] and direct “sequence [DS] CDMA .In both a unique code has been assigned to each transmitter. This code is a pseudo “noise sequence, meaning that it will appear to be noise unless it is known to be a code. So, the transmission can only be decoded by a receiver that knows the codes used.
Both versions of CDMA require a much higher bandwidth than an uncoded transmission. In FH “ CDMA, the higher bandwidth is divided into small bands, each having the bandwidth of uncoded transmission. While sending the data, the used band is changed according to the assigned code. In DS-CDMA the code bits have a higher frequency than the data bits, employing an integer multiplication factor .This factor was chosen as 4 now, based on the available on-chip memory in DSP used. Hence, the original bit stream is multiplied by code bit stream. Since the data is transmitted over a broad frequency band at each fixed point in time ,the receiver can decode the data even if a part of the bandwidth is superimposed by an unrelated transmission. Hence the DS-CDMA has an advantage over FH-CDMA ,because the data can be transmitted even when a particular frequency is not usable due to noise .
As an example of DS “ CDMA , the top section of fig.3 shows the first 12-bits of a pseudo- noise sequence and the middle section of the fig shows the first three bits of a data that are to be coded .The data bits have the duration of 4micro seconds .The CDMA code bits have a duration of one quarter of that of the data bits ,according to the selected spreading factor 4.The data bit stream is multiplied by the code bit stream and the result is shown in bottom part of the figure .
For the present design pseudo “noise sequence has a bit length of 31 bits. This result in 33 orthogonal sequences and subsequently 33 transmitters to be accommodated within transmission system. For installations requiring more transmitters, longer code sequences need to be used. This can be readily implemented, as the modification is only needed at the receiver side.
CDMA provides multiple access communication capabilities. In CDMA each user is provided with an individual pseudo “noise code .It has near -in far “end property.
4. TRANSMITTER DESIGN
Although several wireless transmitters are commercially available on the market, they were not appropriate for the intended bearing condition monitoring application because of their fixed design geometry and packaging forms it is difficult for them to be structurally adapted and integrated into the bearing system to be monitored. Second they contain a receiver circuitry to allow for fault duplex data and communication and error correction. However, such a circuitry requires significantly increased power supply and additional space to physically accommodate the extra components. Given RF transmission alone would account for over 30% of power consumption of entire circuitry, and the receiver section drains more power due to higher number of components, commercial transmitters were not energy optimized in view of power consumption and battery life, and were therefore not appropriate for this.
The digital data transmitter developed in this study transmits a data stream at 2-s intervals, with bandwidth up to 3KHz. A 12-bit resolution is needed to provide proper data resolution. To allow for a generic solution, the analog to digital converter TLV2543 was selected, which provides an interface connection for up to 11 sensors .It has power down mode CDMA encoding and QPSK modulation for data transmission are done by DSP TMS320VC549 [low power consumption and power saving modes]. Sampling every 2-s long data interval with a sampling rate of 6KHz and 12-bit resolution would result 72Kbits/S. To store this amount of data on to the DSP on-chip memory with a 16-bit data width without wasting memory space ,the 12-bit data samples were stored as direct bit strings without any 4-bit prefix .Thus every four 12-bit samples were stored in three DSP memory locations of 16-bit width each ,occupying9Kwords. For CDMA coding there is a total memory need of 36Kwords.
To ensure data transmission integrity over a wireless link, a checksum is added to data sample transmitted ie, recalculated at receiver end .If a mismatch is identified, a retransmission is requested from transmitted. Since transmitter does not include a receiving section on its circuitry it cannot perform inquiry based data retransmission. To improve transmission integrity, each data sample is transmitted three times consecutively. On receiver these transmissions are subsequently compared with each other for consistency. For enhanced transmission integrity, a signal-filtering algorithm can be implemented on the receiver end.
To enable consecutive data transmission, all the data needed must be stored in DSP memory at the same time. As DSP has only 32K word memory space which must hold data, a loop-up table for data modulation and provide memory mapped registers. It is necessary to compress data. This was accomplished by using Lampel “Ziv 1977 data compression algorithm. The algorithm uses dictionary based compression that recognizes repeating patterns in the data and places these patterns in to a dictionary. Compression may vary between 1 to 90%. As the signal contain repetitive periodic waveforms a compression ratio of 50% is expected. The compressed data are CDMA coded and QPSK modulated. For data modulation, the coded bits are sent alternately into two streams. This process is simplified and accelerated by a look-up table. The final sine wave output is stored in another look-up table.Based on phase shift an offset for access to sine look-up table is then selected. In fig 4 a flow chart of data transmission sequence is shown.
The frequency of sine input to RF interface is 78.125KHz. The sine wave is approximated with 16 samples per wave by DSP. The DAC requires 16bit data transmission for each sample to be converted. So a clock frequency of 1.25 MHz is needed for serial communication between DSP and DAC. To convert this signal to carrier frequency a mixer has been implemented which uses the carrier from a local oscillator. The up-converted signal is subsequently transmitted into air via antenna. The RF interface thus was implemented using minimum number of discrete components. No filters were used at base band and for RF frequency to reduce circuit complexity and power consumption. To further minimize CDMA spreading and QPSK modulation were implemented through software.
To ensure program safety for data compression the program is stored in flash memory chip AM29LV400. The flash memory stores sine wave table and pseudo-noise sequence need not be recalculated. In fig 5, the principle design of a transmitter is shown.
The design has been prototyped on a conventional PCB using SMT. The overall size of transmitter using discrete components and SMT packaging, measured about 3.65.8 cm, which is less than half the size of a credit card fig 7. The complete system is shown in fig 8.
Figure 7 Dimensions of Transmitter Board Figure 8 The Complete Transducer system
5. SIMULATION
A random bit generator was employed as data source. These data were first mapped to +1 and-1 to be multiplied by the CDMA code bits. The CDMA codebits were programmed in a repetitive sequence block. The two bit streams were multiplied in next step. The resulting coded data stream is split up into two branches, with even numbered bits in upper branch , odd numbered bits in lower branch. Each branch was then modulated by a carrier wave of 2.4GHz. As the two waveforms have a phase shift of 90degrees to each other, the upper branch is represented by cosine wave and lower sine wave. Both branches were then added together.For simulation RF transmission and receiving sections were assumed to be without losses. The transmission path was modeled to simulate the time delayed transmissions, superpositions, and interface within an enclosed machine environment.
The received signal was demodulated on the receiver end. The demodulation causes a time delay of one-bit duration.As time delay is known in simulation, decoding of CDMA spreading was correlated to coding bit sequence.
6. EXPERIMENTAL EVALUATION
The designed system was experimentally evaluated in two aspects. First the signal strength of the transmitted signal with respect to the transmission distance was investigated to establish the maximum possible transmission range. Second actual power consumption of transmitter was measured.
Experiments have been conducted in lab environment .The signal strength was measured for a transmitter “receiver distance that varied from 15 to 300cm. Two test configurations were used :
1) Transmitter unobstructed on a workbench
2) Transmitter enclosed within a metal casing.
In the latter the casing has a nonmetallic opening of 2*30cm that was covered by a plastic plate of 3cm. The opening was not directed towards the receiver. This situation can be considered as a typical arrangement for a transmitter embedded within a machine environment. In fig 9 both the open space and shielded transmission have shown.
Increasing the distance further will not cause any appreciable signal strength decrease in both cases. From the signal trend analysis it is concluded that the current design is able to ensure reliable transmission within a few tens of meters,which is acceptable for many practical applications. As for the power consumption it is estimated that the transmitter should be able to operate on a set of batteries for a minimum period of three months. The power consumption of transmitter in power-down mode was measured as 4mA. When actively transmitting at 100MHz,the total power consumption of transmitter was 150mA. Reducing clock frequency and transmission speed will reduce power consumption, but lead to prolonged transmission time. For a full cycle of measurements the active transmission period for transmitter is less than 90s. If10 data transmissions were scheduled for each working day of 8hours, atotal 7.5 h active on time would be needed for a period of three months. Such batteries are commercially available in small packages suited for machine embedded applications. For applications requiring extremely small space, the RF amplifier in the present design can be replaced by other low power models at the price of reduced transmission range.
7. ADVANTAGES
¢ Power tuning
¢ Voice quality.
¢ User density.
¢ Cost at a suitable level ie, less cost.
¢ Wireless increases utility and accessibility.
¢ Increased mobility and scalability: more portable, half the size of credit card.
¢ Extended range with CDMA : it allows for multiple transmitters to share the same frequency band, as the receiver can distinguish and identify every specific transmitter by the code, ie, unique.
¢ Since the energy of transmitted data is spread over a broader frequency band than required, even if a portion of frequency band is distorted due to noise, only a part of energy on transmitted data is lost. So the entire data can be reconstructed correctly.
¢ Software implementation allows further modification.
¢ CDMA allows multiple transmitters to share the same frequency band.
8. DISADVANTAGES
¢ When transmitted is embedded with in a machinery the signal strength decreases as distance increases than transmitter in free space.
¢ Ensure reliable transmission only into few lines of meters. It can be increased by designing efficient amplifiers.
¢ Reducing clock frequency and transmission speed for reduced power consumption will lead to a prolonged transmission time.
¢ It reported initial difficulty in market introduction.
¢ WCDMA has higher data speed than CDMA.
9. FUTURE SCOPE
¢ Can be used in mobile communication with a speed up to 2mbps for voice, video data and image transmission with WCDMA.
10. APPLICATIONS
¢ Design of a wide range of electronic instruments such as data loggers, data acquisition cards, hand-held metering devices
¢ Machine health monitoring to machine components
¢ Systems that are difficult to access or not suitable for wired sensor data acquisition.
¢ A rice sized chip called Verichip embedded in body ,made of biocompatible materials, stores entire medical history of implantee.
11. CONCLUSION
A compact, low power digital wireless data transmitter based on CDMA coding technique has been designed, simulated, prototyped and experimentally tested .The design demonstrates the feasibility of of employing a sophisticated transmission scheme in an embedded sensor for machine health monitoring. Focusing on constraints and power efficiency during the design phase.
Has reduced the number of components to a minimum. All the employed components of prototype are available on IC dies and can be integrated to a multichip module that fits on a centimeter sized substrate. The software implementation of CDMA coding allowed for easy extensibility of present design.More systematic experiments will be conducted in future to investigate the utility of such an integrated sensing approach for condition monitoring of bearing in a realistic machine set up.
12. REFERENCES
1. Robert. X . Gao, Philipp Hunerberg- design of a CDMA-Based wireless Data Transmitter for Embedded serving. IEEE TRANSACTION ON INSTRUMENTATION AND MEASSUREMENT Vol:5 p.p. 1259, Dec. 2002
2. Halsall.F,â„¢ Wireless Local Area Networks, Data communications , Computer networks and open systems. p.p. 317-334
3. Thomasi, Digital Communication Electronic Common Systems p.p. 480- 489
4. EFY, Breakthrough in embedded systems, Electronics for you , NOVâ„¢02, pp 65- 69.
5. WWW.TECHONLINE .COM
6. WWW.GOOGLE .COM
7. Dr.Kamilo Feher, Wireless digital communications, Upper saddle river, NJ
rentice Hall, 2002, pp 297-307CONTENTS
1. INTRODUCTION
2. SIGNAL MODULATION TECHNIQUES
A. A.S.K
B. F.S.K
C. CDMA
3. MULTIPLE ACCESS METHODS.
A. FDMA
B. TDMA
C. CDMA
4. TRANSMITTER DESIGN
5. SIMULATION
6. EXPERIMENTAL EVALUATION
7. ADVANTAGES
8. DISADVANTAGES
9. FUTURE SCOPE
10. APPLICATIONS
11. CONCLUSION
12. REFERENCES
ACKNOWLEDGEMENT
I extend my sincere gratitude towards Prof . P.Sukumaran Head of Department for giving us his invaluable knowledge and wonderful technical guidance
I express my thanks to Mr. Muhammed kutty our group tutor and also to our staff advisor Ms. Biji Paul and Mr. Santhosh Kumar for their kind co-operation and guidance for preparing and presenting this seminars.
I also thank all the other faculty members of AEI department and my friends for their help and support.
