USE OF INDUCTIVE PROXIMITY SENSORS IN VITAL SIGNAL APPLICATIONS

Transcription

1 Union Switch & Signal Inc., an Ansaldo Signal company Technology Drive, Pittsburgh, PA 15219? 645 Russell Street, Batesburg, SC USE OF INDUCTIVE PROXIMITY SENSORS IN VITAL SIGNAL APPLICATIONS Copyright 2001, Union Switch & Signal Inc.

2 Table of Contents 1 INTRODUCTION PURPOSE US E OF INDUCTIVE PROXIMITY SENSORS IN VITAL SIGNAL APPLICATIONS...4 List of Figures Figure 1 Typical Proximity Sensor Configuration...5 Figure 2 NAMUR Proximity Sensor Current versus Displacement...6 Figure 3 Multiple Sensor Status and Identification Technique...8

3 1 INTRODUCTION 11, This paper is being presented to the AREMA CNS Technical Conference on September 1.1 Purpose This paper describes a technique of using inductive proximity sensors to determine the position of points and mechanical locking of a switch machine to insure the safe passage of trains. The technique involves multiple sensors and the way in which the status "ON" can be distinguished from "shorted" and the status "OFF" may be distinguished from "open."

4 2 USE OF INDUCTIVE PROXIMITY SENSORS IN VITAL SIGNAL APPLICATIONS Inductive proximity sensors are well suited to the railroad environment. They are hermetically sealed, operate over an extended temperature range and are ultra-reliable. The simplified diagram below, Figure 1., reveals the inductive proximity sensor in the most widely used form. An open-ended ferrite core is an element in a high-frequency oscillator. Normally, the oscillation is at maximum amplitude, but when a metal target is brought within sensing range, eddy currents induced in the target dampen the oscillation. The reduced amplitude is detected, turning on the output transistor and causing the self-contained LED to illuminate. A relay connected to the output will energize when the sensor is in proximity to the target, providing a convenient and easy detection means. Unfortunately, a shorted output transistor will give a false output, indicating close proximity to the target. Since the sensors are very reliable, many applications that are not safety critical are well served by the most common form of a proximity sensor.

5 TYPICAL PROXIMITY SENSOR CONFIGURATION (+) Relay T A R G E T OSCILLATOR LED Out High Frequency Magnetic Field couples with target and induces eddy currents Figure 1 (-) Figure 1. Typical Proximity Sensor Configuration For railway signaling applications, where the safe passage of trains is directly linked to the ON / OFF status of a proximity sensor, a more secure approach is demanded. For this purpose, the NAMUR sensor was selected; it is the simplest form of a proximity sensor. It is a two-wire device characterized by a current-to-displacement relationship as shown in Figure 2. One may think of the NAMUR sensor as a variable resistor controlled by the influence of its proximity to a metal target. In the ON state (near the target) the resistance is about eight times greater than when away from the target. In the nominal detection zone, the resistance changes abruptly, enhancing the precision to which a target may be detected.

6 SHORTED Nominal Detection Range OFF CURRENT INDETERMINATE ON OPEN Figure 2 DISPLACEMENT NAMUR Proximity Sensor Current verses Displacement Figure 2. NAMUR Proximity Sensor Current versus Displacement In the picture below, a typical NAMUR sensor is shown, along with internals before it is potted in a threaded stainless steel barrel. One might realize the simplicity of the device judging from the number of components on the PC board. Only two transistors are used, one for the oscillator and a second for temperature compensation.

7 In a switch machine application, four sensors are deployed. Two are used for point detection and two are used to insure mechanical locking. Each point detector is paired with a corresponding lock detector. An indication is delivered when one pair is proven ON and the second pair is proven OFF. If any sensor is detected in a state other than ON or OFF, no output is delivered. The process is inherently superior to cycle checking, which is only capable of insuring that that status changes when the machine is cycled, but does not guard against the failure of a sensor to mask its state in the static condition. With multiple sensors, the status of a sensor is equally as important as its identification. Figure 3. illustrates the manner in which each sensor is identified and its status determined. This illustration uses only two sensors for simplicity of presentation. It should be recognized that identification of the specific sensor is equally important to determination of its status. For that reason, a signature voltage is assigned to each sensor. With the deployment of four sensors, each is identified by a specific signature voltage.

8 R1 V1 Signature Voltage 1 D/A Tuned Amplifier (F1) Sensor Power SW1 R2 Sensor 1 SW2 Sensor 2 R5 R3 V2 Signature Voltage 2 A/D Microprocessor MUX Tuned Amplifier (F2) Output 1 Output 2 R4 Figure 3 Multiple Sensor Status and Identification Technique Figure 3. Multiple Sensor Status and Identification Technique Each sensor is illustrated as a variable resistor with a series diode, consistent with oneway conduction of current. A microprocessor is deployed with peripheral hardware to complete the system. Diagnostic tests are performed to ensure consistent Analog-to-Digital conversion so that the identification and status of each sensor can be determined. If sensor 1 is identified as ON and sensor 2 as OFF, the processor delivers a frequency F1. Frequency F1 is subsequently selectively power amplified to deliver a DC voltage (output 1). Similarly, output 2 is delivered if sensor 1 is OFF and sensor 2 is ON. Reference is made back to Figure 2. The ON and OFF status of the sensor are narrowly defined as illustrated. The possible range of current is subdivided into five bands. Shorted, open and indeterminate are clearly differentiated from ON and OFF. Such a method considerably reduces the possibility of component failure within the sensor. Thus, the true status of the sensor is indicated absolutely. This technique is inherently superior to simply determining ON or OFF

9 in a cycle-checking sequence. Once the ON state is determined as following narrowly in a current band, component failures within the device cannot mask the ON state as the result of proximity to the target. There are three main time-scheduled tasks performed by the microprocessor. The processor tests linearity and convergence of the A/D by stepping through the entire range of values. In step 2, sensor 1 is identified and its status determined. In step 3, sensor 2 is identified and its status determined. A multiplexer is used to control switches 1 and 2. When SW1 is activated, a voltage is applied to sensor 1; the voltage across R1 identifies sensor 1 as activated. The subsequent voltage read across R5 is then used to determine status. In step 3, sensor 2 is activated and identified from the signature voltage across R4, and its status is determined from the voltage read across R5. The dynamic step-by-step diagnostic testing develops either frequency F1, F2 or neither frequency depending on preset logic residing in the microprocessor. Depending on the status of the two sensors, either output 1, output 2 or no output will be activated, corresponding to the indication output of a switch machine. In the deployment of proximity sensors for switch machine status proving, the technique illustrated in this paper is shown to be consistent with the safety demands of railway signaling applications. Status of a sensor is determined by its analog signature, which enables the distinction between ON and shorted or OFF and open.