Monthly Archives: June 2014

BIOIMPEDANCE

As an introduction to electrical impedance and conduc­tance in biology, a review of the relevant terminology is given and the scope of the discipline is presented. The area of bioimpedance is broad, including, for example, impedance cardiography, electrode impedance, impedance spectroscopy, intraluminal conductance, and impedance to­mography. The field of bioimpedance deals with the elec­trical conduction properties of biological materials as a response to the injection of current. It has been known for more than two centuries that biological structures dis­play the phenomenon of electrical conduction. Later, it was found that the precise electrical properties of tissues depend on their cellular composition and their coupling. These characteristics imply that the voltage changes at a particular site may provide valuable information regard­ing the biological materials and processes concerned.

However, to date, our understanding of the electrical impedance of biological tissues and their changes, as far as they are associated with physiological activity, is still limited. This article discusses the application of electrical impedance in medicine. Briefly, bioimpedance can be used to quantitate extracellular fluid, to assess volume changes, and as an imaging tool similar to ultrasonography. Reviews can be found in (1-3), and (4).

In order to meet the requirements of specific applica­tions, electrodes for delivering or recording electrical po­tentials in biological structures appear in a variety of ma­terials, sizes, and shapes. The interface between electrode and tissue has been studied extensively (5, 6).

ADVANCED DESIGN

In this section, topics are discussed that represent potential performance improvements to future designs. However, none of these are new innovations. The concepts have been known and understood for many years. As technology evolves, though, these techniques become more and more cost effective and attractive to the designer of advanced systems.

Rc Delay Oscillator (Phase Shift Oscillator)

The circuit shown in Fig. 6 is known as the RC (resistor – capacitor) delay oscillator. Three stages of RC delay are used in the feedback path. If the amplifier is inverting, then the RC network needs to produce 180° of phase shift at the frequency of oscillation. Stated differently, the oscil­lation will occur at that frequency at which the feedback has a phase delay of 180° and the gain is greater than one. Other possibilities to realize the phase delays are to RL (resistor-inductor) delay or lossy transmission line delay. A transmission line is just a distributed RLC delay ele­ment. LC delay elements and lossless transmission lines will be covered next.

LC-Tuned Oscillators

When a tuned circuit is used to define the oscillation fre­quency, it is desirable to preserve its frequency selective properties Q. To prevent loading a resonator by the input of an amplifier, a transformer is necessary. Several types of impedance transformation are possible by using cou­pled inductors, inductive division, or capacitive division. Another possibility is to use another amplifier as a buffer: the input impedance of the buffer is high, and the loading on the resonator is light.

Figure 7 shows two ways of looking at how an amplifier can be connected across a tank circuit using capacitive di­vision of voltage. The arrangement shown on the right is known as the Colpitts circuit. Capacitors C1 and C2 in se­ries form part of the resonator, and the input loading by the amplifier is reduced by their ratio. The circuit on the left uses the capacitors as a way of getting 180° of phase shift at resonance, while internally completing the loop with an ad­ditional 180° resulting from signal inversion. However, the two capacitors are still connected in series to each other and across the inductor, and voltage division takes place across this tank. The advantage of capacitive division over inductive division is that it is easier to realize in an in­tegrated circuit. Split inductors or transformers are more complicated to build than simple inductors.

V V V

The arrangements shown in Fig. 8 are known as Hart­ley and Armstrong oscillators, respectively. The feedback in the Hartley circuit is provided by inductive division, using two inductors or a single inductor with a tap connection. Inductive division, using two inductors or a single inductor with a tap connection, provides the feedback in the Hart­ley circuit. The input loading is therefore decreased. The Armstrong circuit uses two inductors, L1 and L2, with in­ductive coupling between them. This may be attractive if the objective is to keep the input and output biases isolated from each other. Moreover, the amount of coupling between the coils can be varied without changing the values of the inductors.

A single transistor implementation of the capacitively divided tank circuit is depicted in Fig. 9. This circuit is simple because the base bias is provided through the in­ductor. The value of the collector load resistor needs to be large enough to prevent decreasing the Q of the resonant circuit. The coil and the series connection of the two capaci­tors make up the resonant circuit. This arrangement can be modified by placing a capacitor in series with the coil (and providing base bias). The series resonant frequency of the coil and the added capacitor can dominate the oscillation frequency, making the circuit less sensitive to variations in the transistor and in the other two capacitors. This modi­fication is known as the Clapp oscillator.

The dc-coupled LC oscillator in Fig. 10 is based on a dif­ferential amplifier. It has a single-ended tank and a single­ended output. Its output is directly coupled to its nonin­verting input, and therefore the feedback needs to provide 0° of phase delay because 360° would be impossible with the circuit shown. At resonance, the LC tank presents a purely resistive load to the collector output, and therefore it has no phase delay. Any internal delay through the device will need to be canceled by adjusting the oscillation frequency slightly out of resonance. The frequency of this oscillator would normally be controlled by using a variable capacitor. Using a variable capacitor would normally control the fre­quency of this oscillator. A small change in frequency can also be obtained by changing the bias on the differential amplifier.

Tuned oscillators may also use transmission lines, cavi­ties, dielectric resonators, crystals, or YIG resonators to de­termine their operating frequency. These resonators have equivalent circuits of very high Q. For example, because of the interaction between voltage and mechanical resonance, piezoelectric quartz crystals are much more stable and of higher Q than LC resonators.

Electrocutaneous Stimulation

The visual and auditory prostheses described above used elec­trical stimulation to activate portions of the nervous system devoted to these functions, and therefore directly produce the sensations of sight or sound. In some conditions, however, damage to the nervous system is such that direct stimulation of the sensory neurons is either not possible or does not lead to sensory perception. For example, blindness caused by dam­age of the visual cortex cannot be addressed by stimulation of this compromised structure. In congenital blindness (where vision was never present), the usual development of the vi­sual cortex may not occur and electrical stimuli applied there may not evoke sensations that can be interpreted in a visual manner. It may also be advantageous to provide an individual with information about the function of an artificial device, such as the grip force produced by a myoelectric artificial arm or the output commands from a hand-grasp neural prosthesis. In such cases, electrical (or mechanical) stimulation of tactile sensors in the skin has been investigated as a means to con­vey information about a different sensory modality, an ap­proach called sensory substitution (79,80). Electrocutaneous stimulation (i. e., electrical stimulation of tactile sensors in the skin) in an area of the skin with intact sensation can be mod­ulated by the variable of interest (e. g., grasp force) so that the user interprets the stimulation in terms of this variable rather than as tactile information. The information can be coded through single electrodes as changes in stimulus ampli­tude, stimulus timing, or both. Multiple electrodes can be ac­tivated progressively as the variable of interest changes, or arrays of electrodes can be used to provide information on inherently multidimensional modalities such as audition and vision.

Natural perception of tactile stimuli depend upon individ­ual receptor properties, how the stimuli are spatially distrib­uted across the skin, and how these stimuli change with time. Current systems for electrocutaneous stimulation do not acti­vate receptors within this normal context, so individuals us­ing this approach must learn to interpret the tactile informa­tion provided in terms of the modality of interest. This has proven to be problematic, and electrocutaneous stimulation has been successfully applied only to a few problems. Blamey and Clark (81) used electrocutaneous stimulation to provide auditory information to profoundly deaf individuals. Sabolich and Ortega (82) used electrocutaneous stimulation to provide center of pressure (a variable related to standing balance) feedback to individuals with artificial lower limbs. The Free­Hand hand-grasp neural prosthesis described previously is implemented with one stimulus channel, providing an electro – cutaneous signal related to the user command signal.

Future work will likely focus upon the factors that have limited success to date. To maximize the transformation of tactile information into the modality being restored, stimulat­ing electrodes must provide more consistent and repeatable inputs to the tactile system, perhaps by implantation. More closely spaced electrodes may allow the tactile system to be activated in a more natural spatial manner. Electrodes and stimulus parameters also need to be optimized to reduce the sensation of pain that often accompanies the tactile sensation produced by electrocutaneous stimulation.

Radar Frequency Interference

There are a large number of radars in operation worldwide. It is likely that several of these systems may be located in close proximity. Because the one-way propagation loss be­tween two systems falls off at only R2, one radar may detect the direct radiation from another. The result is termed radar frequency interference (RFI).

Radar frequency interference may be categorized into two classes. The first class involves two or more radars of the same type. These are designed for the same function, by the same design team, and produced by the same manufacturer. An example is a fleet of ships where each ship carries its own search radar. The second class involves radars of dissimilar type. These have different functions and, probably, have been produced by different companies. An example might be an air­port installation using one high-power general surveillance radar and one or more lower power radars to monitor take­offs and landings.

To minimize RFI, the federal government issues licenses, establishes regulations, and allocates frequency bands to ev­ery radio, television, and radar system deployed within the United States. Other nations have similar functions. Control is extended to all radiating systems including experimental and developmental models.

When two radars in close proximity are of the same type, the designer has some control over RFI. The first defense is channelization. The radar system may be allocated, for exam­ple, a total band of 1000 MHz but its instantaneous radiated bandwidth might be only 10 MHz. Assignment of separate channels spaced by 10 MHz to each particular radar would allow 100 systems to operate simultaneously without any two being in the same channel. Unfortunately, channelization does not completely solve the RFI problem. The radiated spec­trum of a rectangular pulse has very significant energy many bandwidths removed from the carrier frequency. This energy can enter the victim receiver and cause interference.

A second defense is PRF diversity. When two radars oper­ate at the same PRF, the received pulses from the interferer integrate normally and, probably, create a detection. When two different PRFs are used, the integration is negated and the RFI effect is reduced by a factor equal to the integration length. This is not a total panacea. Selection of PRF is a deli­cate process involving questions of blind ranges, blind veloci­ties, and range ambiguity resolution. In a given application, there may not be a sufficient band of usable PRFs to serve multiple radars. Moreover, diversity may even compound the problem. If an interference pulse is sufficiently strong, it may be detectable without integration. This pulse will migrate over many range cells and produce multiple detections. This could result in a swarm of false targets that might overcome the correlation capability of the system computer.

Neither of the techniques discussed above is applicable to the class involving radars of different types. In this case, the only viable defense appears to be asynchronous pulse detec­tion (APD). In the APD technique, successive pulses in a given range cell are compared. If the later pulse is much larger than the earlier pulse, it is edited from the data stream and replaced by either the earlier sample or a random num­ber. Of course, this approach works only when two different PRFs are involved. Asynchronous pulse detection can be very effective in minimizing RFI. However, depending upon the thresholding scheme employed, it can cause degradation in the detectability of real targets.