AMAZON multi-meters discounts AMAZON oscilloscope discounts
1. Introduction
Within instrumentation there is often a need for telemetry in order to transmit data or information between two geographical locations. The transmission may be required to enable centralized supervisory data logging, signal processing, or control to be exercised in large-scale systems which employ distributed data logging or control subsystems. In a chemical plant or power station these subsystems may be spread over a wide area. Telemetry may also be required for systems which are remote or inaccessible, such as a spacecraft, a satellite, or an unmanned buoy in the middle of the ocean.
It can be used to transmit information from the rotating sections of an electrical machine without the need for slip rings.
By using telemetry-sensitive signal processing and recording, an apparatus can be physically remote from hazardous and aggressive environments and can be operated in more closely monitored and controlled conditions.
Telemetry has traditionally been provided by either pneumatic or electrical transmission. Pneumatic transmission, as shown in FIG. 1, has been used extensively in process instrumentation and control. The measured quantity (pressure, level, temperature, etc.) is converted to a pneumatic pressure, the standard signal ranges being 20-100 kPa gauge pressure (3-15 lb/in^2/g) and 20-180 kPa (3-27 lb/in^2/g). The lower limit of pressure provides a live zero for the instrument which enables line breaks to be detected, eases instrument calibration and checking, and provides for improved dynamic response since, when venting to atmospheric pressure, there is still sufficient driving pressure at 20 kPa. The pneumatic signals can be transmitted over distances up to 300 m in 6.35 mm or 9.5 mm od plastic or metal tubing to a pneumatic indicator, recorder, or controller. Return signals for control purposes are transmitted from the control element. The distance is limited by the speed of response, which quadruples with doubling the distance. Pneumatic instrumentation generally is covered at greater length in Section 31.
FIG. 1 Pneumatic transmission.
Pneumatic instruments are intrinsically safe, and can therefore be used in hazardous areas. They provide protection against electrical power failure, since systems employing air storage or turbine-driven compressors can continue to provide measurement and control during power failure.
Pneumatic signals also directly interface with control valves which are pneumatically operated and thus do not require the electrical/pneumatic converters required by electrical telemetry systems, although they do suffer from the difficulty of being difficult to interface to data loggers. Pneumatic transmission systems require a dry, regulated air supply. Condensed moisture in the pipework at subzero temperatures or small solid contaminants can block the small passages within pneumatic instruments and cause loss of accuracy and failure. Further details of pneumatic transmission and instrumentation can be found in Bentley (1983) and Warnock (1985).
Increasingly, telemetry in instrumentation is being under taken using electrical, radio frequency, microwave, or optical fiber techniques. The communication channels used include transmission lines employing two or more conductors which may be a twisted pair, a coaxial cable, or a telephone line physically connecting the two sites; radio frequency (rf) or microwave links which allow the communication of data by modulation of an rf or microwave carrier; and optical links in which the data are transmitted as a modulation of light down a fiber-optic cable. All of these techniques employ some portion of the electromagnetic spectrum, as shown in FIG. 2.
FIG. 2 Electromagnetic spectrum.
FIG. 3 Telemetry system.
FIG. 3 shows a complete telemetry system. Signal conditioning in the form of amplification and filtering normalizes the outputs from different transducers and restricts their bandwidths to those available on the communication channel. Transmission systems can employ voltage, current, position, pulse, or frequency techniques in order to transmit analog or digital data. Direct transmission of analog signals as voltage, current, or position requires a physical connection between the two points in the form of two or more wires and cannot be used over the telephone network. Pulse and frequency telemetry can be used for transmission over both direct links and also for telephone, rf, microwave, and optical links. Multiplexing either on a time or frequency basis enables more than one signal to be transmitted over the same channel. In pulse operation the data are encoded as the amplitude, duration, or position of the pulse or in a digital form.
Transmission may be as a baseband signal or as an amplitude, frequency, or phase modulation of a carrier wave.
In the transmission of digital signals the information capacity of the channel is limited by the available band width, the power level, and the noise present on the channel.
The Shannon-Hartley theorem states that the information capacity, C, in bits/s (bps) for a channel having a bandwidth B Hz and additative Gaussian band-limited white noise is given by:
C=B + log2 (1 + S/N)
where S is the average signal power at the output of the channel and N is the noise power at the output of the channel.
This capacity represents the upper limit at which data can be reliably transmitted over a particular channel. In general, because the channel does not have the ideal gain and phase characteristics required by the theorem and also because it would not be practical to construct the elaborate coding and decoding arrangements necessary to come close to the ideal, the capacity of the channel is significantly below the theoretical limit.
Channel bandwidth limitations also give rise to bit rate limitations in digital data transmission because of intersymbol interference (ISI), in which the response of the channel to one digital signal interferes with the response to the next. The impulse response of a channel having a limited bandwidth of B Hz is shown in FIG. 4(a). The response has zeros separated by 1/2B s. Thus for a second impulse transmitted across the channel at a time 1/2B s later there will be no iSi from the first impulse. This is shown in FIG. 4(b). The maximum data rate for the channel such that no iSi occurs is thus 2B bps. This is known as the Nyquist rate.
FIG. 4(c) shows the effect of transmitting data at a rate in excess of the Nyquist rate.
2. Communication channels
2.1 Transmission lines
Transmission lines are used to guide electromagnetic waves, and in instrumentation these commonly take the form of a twisted pair, a coaxial cable, or a telephone line. The primary constants of such lines in terms of their resistance, leakage conductance, inductance, and capacitance are distributed as shown in FIG. 5. At low frequencies, generally below 100 kHz, a medium-length line may be represented by the circuit shown in FIG. 6, where RL is the resistance of the wire and CL is the lumped capacitance of the line. The line thus acts as a low pass filter. The frequency response can be extended by loading the line with regularly placed lumped inductances.
Transmission lines are characterized by three secondary constants. These are the characteristic impedance, Z0; the attenuation, a, per unit length of line which is usually expressed in dB/unit length; and the phase shift, b, which is measured in radians/unit length. The values of Z0, a, b are related to the primary line constants by:
where R is the resistance per unit length, G is the leakage conductance per unit length, C is the capacitance per unit length, and L is the inductance per unit length.
FIG. 4 (a) Impulse response of a bandlimited channel; (b) impulse responses
delayed by 1/2B s; (c) impulse responses delayed by less than 1/2B s.
FIG. 5 Distributed primary constants of a transmission line.
It is necessary to terminate transmission lines with their characteristic impedance if reflection or signal echo is to be avoided. The magnitude of the reflection for a line of characteristic impedance Z0 terminated with an impedance ZT is measured by the reflection coefficient, ρ, given by:
ρ = (ZT - Z0) / (ZT + Z0)
Twisted pairs are precisely what they say they are, namely, two insulated conductors twisted together. The conductors are generally copper or aluminum, and plastic is often used as the insulating material. The twisting reduces the effect of inductively coupled interference. Typical values of the primary constants for a 22 gauge copper twisted pair are R = 100 O/km, L = 1 mH/km, G = 10-5 S/km and C = 0.05 nF/km. At high frequencies the characteristic impedance of the line is approximately 140 O. Typical values for the attenuation of a twisted pair are 3.4 dB/km at 100 kHz, 14 dB/km at 1 MHz, and 39 dB/km at 10 MHz. The high-frequency limitation for the use of twisted pairs at approximately 1 MHz occurs not so much as a consequence of attenuation but because of crosstalk caused by capacitive coupling between adjacent twisted pairs in a cable.
FIG. 6 Low-frequency lumped approximation for a transmission line.
FIG. 7 Coaxial cable.
FIG. 8 Gain and delay distortion on telephone
lines.
Coaxial cables which are used for data transmission at higher frequencies consist of a central core conductor surrounded by a dielectric material which may be either poly ethylene or air. The construction of such cables is shown in FIG. 7. The outer conductor consists of a solid or braided sheath around the dielectric. In the case of the air dielectric the central core is supported on polyethylene spacers placed uniformly along the line. The outer conductor is usually covered by an insulating coating. The loss at high frequencies in coaxial cable is due to the "skin effect," which forces the current in the central core to flow near to its surface and thus increases the resistance of the conductor. Such cables have a characteristic impedance of between 50 and 75 O. The typical attenuation of a 0.61 cm diameter coaxial cable is 8 dB/100 m at 100 MHz and 25 dB/100 m at 1 GHz.
Trunk telephone cables connecting exchanges consist of bunched twisted conductor pairs. The conductors are insulated with paper or polyethylene, the twisting being used to reduce the crosstalk between adjacent conductor pairs. A bunch of twisted cables is sheathed in plastic, and the whole cable is given mechanical strength by binding with steel wire or tape which is itself sheathed in plastic. At audio frequencies the impedance of the cable is dominated by its capacitance and resistance. This results in an attenuation that is frequency dependent and phase delay distorted, since signals of different frequencies are not transmitted down the cable with the same velocity. Thus a pulse propagated down a cable results in a signal which is not only attenuated (of importance in voice and analog communication) but which is also phase distorted (of importance in digital signal transmission). The degree of phase delay distortion is measured by the group delay db/dO. The bandwidth of telephone cables is restricted at low frequencies by the use of ac amplification in the repeater stations used to boost the signal along the line. Loading is used to improve the high-frequency amplitude response of the line.
This takes the form of lumped inductances which correct the attenuation characteristics of the line. These leave the line with a significant amount of phase delay distortion and give the line attenuation at high frequencies. The usable frequency band of the telephone line is between 300 Hz and 3 kHz. FIG. 8 shows typical amplitude and phase or group delay distortions relative to 800 Hz for a typical leased line and a line to which equalization or conditioning has been applied.
In order to transmit digital information reliably the trans mission equipment has to contend with a transmission loss which may be as high as 30 dB; a limited bandwidth caused by a transmission loss which varies with frequency; group delay variations with frequency; echoes caused by impedance mismatching and hybrid crosstalk; and noise which may be either Gaussian or impulsive noise caused by dial pulses, switching equipment, or lightning strikes. Thus it can be seen that the nature of the telephone line causes particular problems in the transmission of digital data. Devices known as modems (Modulators/dEModulators) are used to transmit digital data along telephone lines. These are considered in Section 9.1.
2.2 Radio Frequency transmission
Radio frequency (rf) transmission is widely used in both civilian and military telemetry and can occur from 3 Hz (which is referred to as very low frequency (VLF)) up to as high as 300 GHz (which is referred to as extremely high frequency (EHF)). The transmission of the signal is by means of line-of-sight propagation, ground or surface wave diffraction, ionospheric reflection or forward scattering (Coates 1982). The transmission of telemetry or data signals is usually undertaken as the amplitude, phase, or frequency modulation of some rf carrier wave. These modulation techniques are described in Section 5. The elements of an rf telemetry system are shown in FIG. 9.
The allocation of frequency bands has been internationally agreed upon under the Radio Regulations of the international Telecommunication Union based in Geneva. These regulations were agreed to in 1959 and revised in 1979 (HMSo, 1980). In the UK the Radio Regulatory division of the department of Trade approves equipment and issues licenses for the users of radio telemetry links. In the United States, the Federal Communications Commission (FCC) serves the same purpose. In other countries, there is an analogous office. For general-purpose low-power telemetry and telecontrol there are four bands which can be used. These are 0-185 kHz and 240-315 kHz, 173.2-173.35 MHz and 458.5-458.8 MHz. For high-power private point systems the allocated frequencies are in the UHF band 450 470 MHz. In addition, systems that use the cellular telephony bands are becoming common.
For medical and biological telemetry there are three classes of equipment. Class i are low-power devices operating between 300 kHz and 30 MHz wholly contained within the body of an animal or human. Class ii is broad-band equipment operating in the band 104.6-105 MHz. Class iii equipment is narrow-band equipment operating in the same frequency band as the Class ii equipment. details of the requirements for rf equipment can be found in the relevant documents cited in the References (HMSo, 1963, 1978, 1979).
2.3 Fiber-optic communication
Increasingly, in data-communication systems there is a move toward the use of optical fibers for the transmission of data.
Detailed design considerations for such systems can be found in Keiser (1983), Wilson and Hawkes (1983), and Senior (1985). As a transmission medium fiber-optic cables offer the following advantages:
1. They are immune to electromagnetic interference.
2. Data can be transmitted at much higher frequencies and with lower losses than twisted pairs or coaxial cables.
Fiber optics can therefore be used for the multiplexing of a large number of signals along one cable with greater distances required between repeater stations.
3. They can provide enhanced safety when operating in hazardous areas.
4. Ground loop problems can be reduced.
5. Since the signal is confined within the fiber by total internal reflection at the interface between the fiber and the cladding, fiber-optic links provide a high degree of data security and little fiber-to-fiber crosstalk.
6. The material of the fiber is very much less likely to be attacked chemically than copper-based systems, and it can be provided with mechanical properties which will make such cables need less maintenance than the equivalent twisted pair or coaxial cable.
7. Fiber-optic cables can offer both weight and size advantages over copper systems.
FIG. 9 RF telemetry system.
FIG. 10 Elements of an optical fiber.
FIG. 11 Propagation down fibers.
2.3.1 Optical Fibers
The elements of an optical fiber, as shown in FIG. 10, are the core material, the cladding, and the buffer coating.
The core material is either plastic or glass. The cladding is a material whose refractive index is less than that of the core.
Total internal reflection at the core/cladding interface con fines the light to travel within the core. Fibers with plastic cores also have plastic cladding. Such fibers exhibit high losses but are widely used for short-distance transmission.
Multicomponent glasses containing a number of oxides are used for all but the lowest-loss fibers, which are usually made from pure silica. In low- and medium-loss fibers the glass core is surrounded by a glass or plastic cladding. The buffer coating is an elastic, abrasion-resistant plastic material which increases the mechanical strength of the fiber and provides it with mechanical isolation from geometrical irregularities, distortions, or roughness of adjacent surfaces which could otherwise cause scattering losses when the fiber is incorporated into cables or supported by other structures.
The numerical aperture (NA) of a fiber is a measure of the maximum core angle for light rays to be reflected down the fiber by total internal reflection.
By Snell's Law:
NA = sin θ √ (μ21 - μ22)
where μ1 is the refractive index of the core material and μ2 is the refractive index of the cladding material.
Fibers have NAs in the region of 0.15-0.4, corresponding to total acceptance angles of between 16 and 46 degrees.
Fibers with higher NA values generally exhibit greater losses and low bandwidth capabilities.
The propagation of light down the fibers is described by Maxwell's equations, the solution of which gives rise to a set of bounded electromagnetic waves called the "modes" of the fiber. Only a discrete number of modes can propagate down the fiber, determined by the particular solution of Maxwell's equation obtained when boundary conditions appropriate to the particular fiber are applied. FIG. 11 shows the propagation down three types of fiber. The larger core radius multimode fibers are either step index or graded index fibers. In the step index fibers there is a step change in the refractive index at the core/cladding interface. The refractive index of the graded index fiber varies across the core of the fiber. Monomode fibers have a small-core radius, which permits the light to travel along only one path in the fiber.
The larger-core radii of multimode fibers make it much easier to launch optical power into the fiber and facilitate the connecting of similar fibers. Power can be launched into such a fiber using light-emitting diodes (LEDs), whereas single-mode fibers must be excited with a laser diode.
Intermodal dispersion occurs in multimode fibers because each of the modes in the fibers travels at a slightly different velocity. An optical pulse launched into a fiber has its energy distributed among all its possible modes, and therefore as it travels down the fiber the dispersion has the effect of spreading the pulse out. Dispersion thus provides a bandwidth limitation on the fiber. This is specified in MHz x km. In graded index fibers the effect of intermodal dispersion is reduced over that in step index fibers because the grading bends the various possible light rays along paths of nominal equal delay. There is no intermodal dispersion in a single-mode fiber, and therefore these are used for the highest-capacity systems. The bandwidth limitation for a plastic clad step index fiber is typically 6-25 MHz x km. Employing graded index plastic-clad fibers this can be increased to the range of 200-400 MHz x km. For monomode fibers the bandwidth limitation is typically 500-1500 MHz x km.
Attenuation within a fiber, which is measured in dB/km, occurs as a consequence of absorption, scattering, and radiative losses of optical energy. Absorption is caused by extrinsic absorption by impurity atoms in the core and intrinsic absorption by the basic constituents of the core material.
One impurity which is of particular importance is the oH (water) ion, and for low-loss materials this is controlled to a concentration of less than 1 ppb. Scattering losses occur as a consequence of microscopic variations in material density or composition, and from structural irregularities or defects introduced during manufacture. Radiative losses occur whenever an optical fiber undergoes a bend having a finite radius of curvature.
Attenuation is a function of optical wavelength. FIG. 12 shows the typical attenuation versus wavelength characteristics of a plastic and a monomode glass fiber. At 0.8 nm the attenuation of the plastic fiber is 350 dB/km and that of the glass fiber is approximately 1 dB/km. The minimum attenuation of the glass fiber is 0.2 dB/km at 1.55 nm. FIG. 13 shows the construction of the light- and medium-duty optical cables.
FIG. 12 Attenuation characteristics of optical fibers.
FIG. 13 Light- and medium-duty optical cables.
FIG. 14 Spectral output from a LED.
FIG. 15 Spectral output from a laser diode.
FIG. 16 LED and p-i-n diode detector for use in a fiber-optic system.
2.3.2 Sources and Detectors
The sources used in optical fiber transmission are LEDs and semiconductor laser diodes. LEDs are capable of launching a power of between 0.1 and 10 mW into the fiber. Such devices have a peak emission frequency in the near infrared, typically between 0.8 and 1.0 nm. FIG. 14 shows the typical spectral output from a LED. Limitations on the transmission rates using LEDs occur as a consequence of rise time, typically between 2 and 10 ns, and chromatic dispersion. This occurs because the refractive index of the core material varies with optical wavelength, and therefore the various spectral components of a given mode will travel at different speeds.
Semiconductor laser diodes can provide significantly higher power, particularly with low duty cycles, with out puts typically in the region of 1 to 100 mW. Because they couple into the fiber more efficiently, they offer a higher electrical to optical efficiency than do LEDs. The lasing action means that the device has a narrower spectral width compared with a LED, typically 2 nm or less, as shown in FIG. 15. Chromatic dispersion is therefore less for laser diodes, which also have a faster rise time, typically 1 ns.
For digital transmissions of below 50 Mbps LEDs require less complex drive circuitry than laser diodes and require no thermal or optical power stabilization.
Both p-i-n (p material-intrinsic-n material) diodes and avalanche photodiodes are used in the detection of the optical signal at the receiver. In the region 0.8-0.9 nm silicon is the main material used in the fabrication of these devices.
The p-i-n diode has a typical responsivity of 0.65 A/W at 0.8 nm. The avalanche photodiode employs avalanche action to provide current gain and therefore higher detector responsivity. The avalanche gain can be 100, although the gain produces additional noise. The sensitivity of the photodetector and receiver system is determined by photodetector noise which occurs as a consequence of the statistical nature of the production of photoelectrons, and bulk and dark surface current, together with the thermal noise in the detector resistor and amplifier. For p-i-n diodes the thermal noise of the resistor and amplifier dominates, whereas with avalanche photodiodes the detector noise dominates.
FIG. 16 shows a LED and p-i-n diode detector for use in a fiber-optic
system.
2.3.3 Fiber-Optic Communication Systems
FIG. 17 shows a complete fiber-optic communications system. In the design of such systems it is necessary to compute the system insertion loss in order that the system can be operated using the minimum transmitter output flux and minimum receiver input sensitivity. In addition to the loss in the cable itself, other sources of insertion loss occur at the connections between the transmitter and the cable and the cable and the receiver; at connectors joining cables; and at points where the cable has been spliced. The losses at these interfaces occur as a consequence of reflections, differences in fiber diameter, NA, and fiber alignment. Directional couplers and star connectors also increase the insertion loss.
FIG. 17 Fiber-optic communication system.
FIG. 18 (a) Frequency-division multiplexing: (b) transmission of three
signals using FDM.
3. Signal multiplexing
In order to enable several signals to be transmitted over the same medium it is necessary to multiplex the signals. There are two forms of multiplexing: frequency-division multiplexing (FDM) and time-division multiplexing (TDM). FDM splits the available bandwidth of the transmission medium into a series of frequency bands and uses each of the frequency bands to transmit one of the signals. TDM splits the transmission into a series of time slots and allocates certain time slots, usually on a cyclical basis, for the transmission of one signal.
The basis of FDM is shown in FIG. 18(a). The bandwidth of the transmission medium fm is split into a series of frequency bands, having a bandwidth fch, each one of which is used to transmit one signal. Between these channels there are frequency bands, having bandwidth fg, called "guard bands," which are used to ensure that there is adequate separation and minimum cross-talk between any two adjacent channels. FIG. 18(b) shows the transmission of three band-limited signals having spectral characteristics as shown, the low-pass filters at the input to the modulators being used to bandlimit the signals. Each of the signals then modulates a carrier. Any form of carrier modulation can be used, although it is desirable to use a modulation which requires minimum bandwidth. The modulation shown in FIG. 18(b) is amplitude modulation (see Section 5).
The individually modulated signals are then summed and transmitted. Bandpass filters after reception are used to separate the channels, by providing attenuation which starts in the guard bands. The signals are then demodulated and smoothed.
TDM is shown schematically in FIG. 19. The multiplexer acts as a switch connecting each of the signals in turn to the transmission channel for a given time. In order to recover the signals in the correct sequence it is necessary to employ a demultiplexer at the receiver or to have some means inherent within the transmitted signal to identify its source. If N signals are continuously multiplexed then each one of them is sampled at a rate of 1/N Hz. They must therefore be band limited to a frequency of 1/2N Hz if the Shannon sampling theorem is not to be violated.
The multiplexer acts as a multi-input-single-output switch, and for electrical signals this can be done by mechanical or electronic switching. For high frequencies electronic multiplexing is employed, with integrated circuit multiplex ers which use CMOS or BiFET technologies.
TDM circuitry is much simpler to implement than FDM circuitry, which requires modulators, band-pass filters, and demodulators for each channel. In TDM only small errors occur as a consequence of circuit non-linearities, whereas phase and amplitude non-linearities have to be kept small in order to limit intermodulation and harmonic distortion in FDM systems. TDM achieves its low channel crosstalk by using a wideband system. At high transmission rates errors occur in TDM systems due to timing jitter, pulse accuracy, and synchronization problems. Further details of FDM and TDM systems can be found in Johnson (1976) and Shanmugan (1979).
FIG. 20 Pulse code modulation.
FIG. 21 Types of pulse code modulation.
FIG. 22 other forms of pulse encoding.
Related Articles -- Top of Page -- Home