Overcoming problems with inductive Sensors

John Francis from Positek explains a new technology that can offer the advantages of a non contact inductive sensor but with the simplicity and cost effectiveness of potentiometers

Sensors based on inductive principles offer the best potential compromise between robustness and performance. LVDT technology has proven itself in the industrial and aerospace markets even though it is based on old technology which has several disadvantages. The coil structures are large with very high turn counts of delicate wire and the overall length is up to three times the travel. This adds to the cost and makes installation difficult.

What is more, to achieve the best performance, the interface electronics unit is often large and bulky, and has to be mounted independently of the sensor. And to achieve good temperature stability, the electronic modules must be matched to the sensors making the installation even more complex.

Designers at Positek set out to develop a new technology which would overcome these difficulties and yet retain the advantages of inductive technology. In addition, they wanted to develop a totally self contained unit allowing the customer to install the sensor and interface it with his control or monitoring system.

The new technology is called PIPS for Positek’s Inductive Position Sensors, and has two derivatives: LIPS, linear position sensors and RIPS, a rotary version. Both versions provide a simple analogue output from a dc supply voltage. A variety of industry standard interfaces has been designed using a modular approach to simplify the construction.

The new PIPS sensors use a primary sensing element which is a variable inductive coil arranged in a simple half bridge. A full bridge resonant circuit is created by coupling two capacitors across the coil. With the circuit driven into resonance by an oscillator, the output is taken from the centre tapping. As the relative inductance of the two arms of the bridge is varied, the peak voltage of the output signal is modulated, (Figure 2). This is a well understood fundamental principle.

The PIPS coil is very different to other inductive sensors in that it uses a coil with a few turns of wire, typically 100 turns. In a linear sensor, the coils are wound on a former with standard enamelled copper wire. In rotary sensors, the coil is produced using standard PCB manufacturing techniques.

The result is that the coil can be very compact and is very robust. It is possible to produce linear sensors with coils of only 1.5mm diameter. With only a few turns, the coil has a low inductance and with appropriate capacitor values, the natural frequency of oscillation of the LC circuit is set around 1MHz to 2MHz. The frequency is not specifically controlled and can vary as the target moves over the coil.

To change the inductance of the coil, a conductive target passes over or through the coil and eddy currents are induced in the target which changes the inductance of the coil. Aluminium and some stainless steels are good examples of suitable materials for influencing the inductance due to their low cost and well defined and consistent electrical properties. The target only needs to be 0.1mm, since the eddy currents flow in the material’s surface.

Any sensor must consistently produce a stable output signal which is representative of the position of the target being measured. This must be achieved over a wide range of environmental and operating conditions. A transducer coil itself is not the only component that must produce a stable and consistent signal. The other important element is the electronic interface which decodes the raw oscillating signal into a signal which can be easily used by the customer.

Consistently decoding a signal which is oscillating at 1MHz is not easy but with PIPS there has been a significant innovation in the interface circuit. The interface circuit produces an analogue signal from the varying amplitude of the signal from the centre of the bridge. The interface measures the amplitude of the output signal by taking a sample at exactly the same point in the cycle each cycle.

The main problem is to ensure that the sample is taken at exactly the same point in the cycle even though the operating frequency varies as the coil inductance varies. To do this, the decode circuit uses two internal servo loops to align the sampling of the output signal and to control the amplitude of the drive signal.

Figure 3 shows the oscillator drive signal and the timing arrangement for the sampling of the signals. Two switches are connected to the drive from the oscillator and controlled so that they take consecutive samples of the drive signal. Switch B closes first after a delay period A which is triggered by the zero crossing of the signal. The sample taken from switch B is fed into a comparator and immediately afterwards the sample taken from switch C is fed into the comparator.

Both switches are closed for the same duration of about 60ns. Any resulting difference between the two samples is fed back to adjust the delay period A. Eventually, the timing of the two switches is such that the sample voltages are equal. At this condition, both switches are equally spaced about the peak of the drive signal.

There is a stable and controlled timing arrangement which is related to the drive signal and which is used to control the timing and sampling of the output switch D. In addition to the timing of the samples, it is important to control the amplitude of the drive signal. This is simply achieved by comparing the sample voltage B with a reference voltage and adjusting the oscillator drive voltage.

The samples from the output switch are fed to an output amplifier and filter stage. The gain and offset of the response characteristic may be adjusted by the customer to suit his application.

The basic interface electronics provides all the control and drive functions from a 5V supply and will produce a voltage output signal from 0 to 5V for the full travel. The interface includes both gain and offset adjustment for calibration and customer adjustment purposes.

Common interface requirements include 0-10V output 610(or5)V and 4-20mA loop powered. These can be catered for by incorporating additional signal conditioning with supply regulation and output amplification in modular assemblies built around the basic interface.

All the PIPS sensors have a two-wire loop powered 4-20mA option, achieved by incorporating a special power supply configuration. The 4-20mA option is useful in large process control applications where cable lengths between controller and sensor are long. Sensors have been designed and built for linear travels of between 1mm and 1500mm and from 1.5mm diameter to 45mm diameter.

Up until recently, there have been few rotary sensors which have provided the kind of price and performance of a potentiometer but without the disadvantage of poor durability. Traditionally the only non-contact alternatives to a potentiometer were RVDTs, resolvers and optical encoders. With RIPS it is now possible to produce a much more cost effective, compact and durable device, (Figure 4).

The coils are produced using standard PCB techniques and the target which affects the inductance of the coils is made up of a pair of semicircular vanes which are mounted either side of the PCB. The coil structure on the PCB is designed to produce optimal linearity over a large angle. At the present time, it is possible to achieve good linearity of 61% over 140 degrees, and it is expected that in the future this will be nearly 360 degrees. Over reduced angles, the linearity is improved.

The vanes are spaced equally from the coil board and are mounted onto the sensor input shaft. In the design, special effort has been made to eliminate costly end thrust bearing arrangements and, in the case of RIPS, the PCB takes all the end thrust. Control of the input shaft is achieved using either plain or sealed ball race bearings.

A variety of linear sensors has been produced too. The flexibility of the coil design and the lack of special target materials allows sensors to be designed into machinery, eliminating several components and simplifying installation. A LIPS sensor can be applied in a brake servo system to measure the piston directly without interfering with the linkage from the pedal to the piston.

Cylinder sensors are designed for use in hydraulic and pneumatic cylinders or where environmental conditions dictate a fully sealed assembly such as internal gearbox controls, (Figure 5). These sensors are designed with a coil wound on a stiff fibreglass or carbon fibre core and sheathed with a high performance polymer which provides fluid resistance. The coil and sheath are epoxy filled to produce a pressure insensitive assembly.

The coil is wound in a single layer with a coarse helix and the pitch of the helix is controlled to give a linear response. In linear sensors, the ratio of the overall length to the travel is important. With LIPS sensors this ratio can be almost 1:1 since it is configured with only half of the bridge circuit active. The active coil is wound over the length of the probe whilst the other half of the bridge is a very short reference coil. In some cases this can be a commercial inductor mounted with the electronics but for optimum performance good material and temperature matching is achieved with a wound reference coil.

Sensors with travels of between 4mm and at least 600mm can be accommodated with this design, and standard levels of linearity of 0.25% produced. By fine tuning the winding profile, good linearity can be achieved over almost the full length of the active coil.

Other sensors use an external coil with an internal pushrod target, (Figure 6). They are provided with an integral bearing assembly and have a short reference coil mounted over the bearing section with the active coil spaced along the majority of the sensor.

The electronics is mounted in the end of the sensor and the whole assembly is mounted in a stainless steel housing tube. The sensors can be produced with travels up to 1500mm and are available with a number of options. The sensor can be supplied with rod end fittings or with body clamps and front flange mountings.

One of the major advantages of the PIPS technology is the very small coil size which is required. Hence, it has the ability to produce extremely small sensors for measuring short travels of up to 10mm. The sensor coil is fully differential and has a diameter of only 2.6mm. The housing for the coil is a ceramic or plastic tube which has an outside diameter of 4mm. The sensors uses an external concentric tube as the target.

In miniature LIPS sensors, (Figure 7), the construction is similar to the cylindrical sensor, except that the coil is constructed in two identical parts. In the mid travel position, the target covers half of each coil. This gives a fully balanced situation which improves temperature stability at this position. The linearity of the sensor is within 1% for a travel of 1mm close to the mid point of the coil. Drift is less than 10um for a temperature change of 50 degreesC.

{{PositekTel: Cheltenham (01242) 820027Enter 530}}

Figure 1: The Brown Brothers ship stabiliser system uses a Positek cylinder LIPS system sensor mounted into the main fin actuator

Figure 2: As the relative inductance of the two arms of the bridge is varied the peak voltage of the output signal is modulated

Figure 3: The oscillator drive signal and the timing arrangementfor the sampling ofthe signals

Figure 4: Construction of the RIPS rotary sensor

Figure 5: Cylinder sensors are designed for use in fluid power applications in hydraulic and pneumatic cylinders

Figure 6: This sensor uses an external coil with an internal pushrod as the target. The sensors are provided with an integral bearing assembly and have a short reference coil mounted over the bearing section with the active coil spaced along the majority of the sensor