Integration of piezoresistive measuring cells into existing applications

Integration of piezoresistive measuring cells into existing applications

The core element of every pressure transmitter is the pressure measurement cell. With piezoresistive pressure transmitters, this equates essentially to the Wheatstone bridge measuring arrangement. The primary pressure measurement takes place here by way of deformations to the strain gauges. This piezoresistive measuring cell can also be integrated into existing applications such as pressure switches or pressure regulators, should this be necessary. Various possibilities exist to this end.

The most common reason for the need to integrate a sensor cell instead of a pressure transmitter into an existing application is a lack of space. In hydraulic valves, for example, there are only a few cubic centimeters of space. The integration of an entire pressure transducer is therefore not usually possible. Because of insufficient space, some users opt to employ an external sensor, which is then flange-mounted to the existing application. This approach, however, is cumbersome and not as optimal as the integration of separate measuring cells into the application.

In the selection of suitable measuring cells for individual applications, the same questions apply by and large as with the selection of an entire pressure transmitter.What needs to be established, among other things, are the pressure range to be measured, the temperature conditions and also the relevant media compatibility. In the employment of piezoresistive measuring cells into existing applications, three further selection criteria can be added: These are the mechanical and electrical considerations for integrating the sensor cells.

The mechanical selection criterion relates to actually building the measuring cells into the relevant application. Depending upon requirements, these possibilities remain open:

  • screw in
  • weld on
  • plug in
  • wedge in

On the electrical side, it must be determined which electronics are used in the application to provide the electrical signal transmission. In some circumstances, it may be that the electronics existing in the application are not equipped for the integration of pressure measurement cells. In this case, an electrical signal conversion would have to be separately integrated.

We now arrive at a real life example: An STS customer wanted to retrofit an existing precision high-pressure control valve for test bench applications with the option of pressure measurement. Since an entire pressure transmitter could not be integrated into the valve, a single pressure measurement cell had to be opted for. The demands here were that it had to display pressures up to 600 bar and it should be designed for a signal output from 0 to 100 mV/V at a supply of 10 V.

The solution selected was a measuring cell with stainless steel pressure port and miniature compensation technology. This could be screwed into the valve body below the already existing cover in a space-saving manner and also shielded from external influences. The assembly height after mounting on the valve body came to under 30 mm (including bending radius of cord strands). Apart from its minimal dimensions, there was one additional feature: The zero position and range were individually adaptable by the user with a potentiometer.

Measuring cell with stainless steel pressure port for implementation on a high-pressure control valve

Consultancy is key

Piezoresistive measuring cells are the core competency at STS. They are fully manufactured in-house, display pressure ranges from 100 mbar to 1,000 bar and are available in the materials of stainless steel, titanium and Hastelloy®. This means that, in principle, they can be employed for almost any conceivable measuring task. In collaboration with our engineers, customers receive an extensive consultancy on the integration of suitable measuring cells into existing applications.

Strain gauges in pressure measurement technology

Strain gauges in pressure measurement technology

Strain gauges are measuring devices that change their electrical resistance through mechanical deformation. They are used in a variety of measuring instruments, which, besides scales and load cells, also includes pressure sensors.

Pressure sensors rely on several physical variables, including inductance, capacitance or piezoelectricity. The most common physical property by which pressure transmitters operate, however, is the electrical resistance that can be observed in the metallic deformation, or piezoresistive effect, of semiconductor strain gauges. The pressure is determined by a mechanical deformation, where strain gauges are attached to an elastic carrier. It is important here that the strain gauges can follow the movements of this carrier. If a pressure acts on the carrier, the deformation arising brings about a change in cross-section of the conductor tracks, which in turn causes a shift in the electrical resistance. It is ultimately this change in electrical resistance that a pressure transducer records and from which the pressure can then be determined.

Figure 1: Strain gauges deform under pressure

The deformation acting upon the conductor will thus cause it to change in length (Δl). Since the volume remains the same, it is the cross-section and thus the resistance R that changes:

ΔR/R = k • Δl/l

The change in resistance (ΔR) is proportional to the change in length (Δl), and the proportionality factor (k) will depend on both the geometry and the material properties. While ‘k’ will be 2 for metallic conductors, it can also be very high in semiconductors. Because of these relatively high ‘k-factors’ for semiconductors, these are more sensitive and can therefore measure even the slightest of pressure changes. Temperature dependency, however, also increases as a result of this.

The change in resistance in metallic strain gauges results from dimensional changes (geometry). In semiconductor strain gauges, however, the change is due to alterations in the crystal structure (piezoresistive effect).

The evaluation of the resistance change triggered by a pressure-induced deformation then takes place via a bridge circuit. For this purpose, the strain gauges are connected up to form a Wheatstone bridge (Figure 2). Two of the strain gauges are placed in a radial direction and two in a tangential one. It is thus so that two become stretched and two become compressed under deformation. In order for temperature effects to be compensated and for the signal to be as linear as possible, it is important that the strain gauges have the exact same resistances and are arranged in an exact geometry.

Figure 2: Bridge circuit

Metallic strain gauges

Among metallic strain gauges, we must differentiate between the foil and thin-film varieties.

Foil strain gauges consist of rolled foil, only a few microns thick. Constantan is normally used as the material here, but Karma and Modco can also be employed, especially if a larger temperature range is needed or the temperatures are below -150 °C. Constantan has a very low ‘k-factor’ of 2.05 and is therefore not very sensitive. Considering this, the material displays a lowered temperature dependency, which is also why it is most often used in foil strain gauges.

Foil strain gauges are more likely to be used in load cells. Often they are not sensitive enough to be pressure transducers, since values of less than one bar cannot be recorded with them. Their temperature range is also relatively limited, and, depending on the version, temperatures of even 80 °C should not be exceeded.

Thin-film strain gauges are produced by the so-called thin-film technique, by, for example, vapor deposition or sputter coating. The manufacturing process is more complex here and also more expensive than for foil gauges. On the other hand, however, a temperature range of 170 °C is possible, with their long-term stability also being very good.

Metallic thin-film strain gauges provide for stable over the longer term, but also quite expensive, measuring instruments. It holds true that the lower the pressures to be detected are, the higher the manufacturing cost will be. Low pressures of less than 6 bar can only be detected at a poor accuracy.

Semiconductor strain gauges

Semiconductor strain gauges operate by the piezoresistive effect. The material used in most cases is silicon. Semiconductor strain gauges tend to be more sensitive than the metallic variety. They are also usually separated from the medium by a separation membrane, with the pressure being passed on via a transfer fluid.

Figure 3: Piezoresistive measuring device

In semiconductor materials, the piezoresistive effect is about fifty times more pronounced than with metallic strain gauges. The semiconductor strain gauges are either glued to a carrier or directly sputter-coated onto it. The latter enables an intense bonding and assures freedom from hysteresis, as well as a resistance to aging and temperature stability. Although the piezoresistive effect is not exclusive to the semiconductor strain gauge, the term “piezoresistive pressure sensor” has come to be used for instruments where the elastic structure deforming under pressure and the resistors are all integrated into one chip. Piezoresistive pressure transducers can be made small in size and (apart from the membrane) without any moving parts. Their production is based on normal semiconductor fabrication methods. At the same time, there is the possibility of integrating the resistors with the elastic membrane deforming under pressure all into one chip and thus produce a full pressure measurement cell in the size of just one chip.

Piezo thin-film strain gauges are attached to a silicon carrier and separated from the carrier by an insulating layer. This increases the manufacturing requirement and thus also the price, but temperature ranges from -30 °C to 200 °C are possible here. Thanks to the highly elastic properties of silicon, only a low hysteresis can be expected with these. It is the high ‘k-factor’ that achieves the high sensitivity, making piezoresistive pressure transmitters the first choice for the smallest of pressure ranges on the mbar scale. In addition, devices of tiny dimension can be produced, which has a positive effect on the scope of potential applications. Also, the long-term stability and EMC compatibility is very good, with the latter, of course, depending upon carrier material. Temperature compensation, however, requires a little more effort, but even this challenge can also be overcome quite easily. You can read more about temperature compensation here.

Thick-film strain gauges are printed onto ceramic or metallic membranes. With a thickness of 20 microns, they are up to 1,000 times thicker than thin-film strain gauges. Because of their low production requirements, these are cheaper in price, but not very stable longer term due to the aging of their thick film.

Summary: The type of strain gauge used has a major influence on the measuring instrument. Factors such as price, accuracy and long-term stability play an important role in choosing the right pressure transmitter. In our experience, pressure transmitters with piezo thin-film strain gauges have proven to be the most efficient, because, thanks to their sensitivity, they can record wide pressure ranges at high accuracy, whilst also exhibiting good long-term stability.

Mechanical simulation prior to demanding pressure measurement projects

Mechanical simulation prior to demanding pressure measurement projects

Engineering methods and modern technologies enable manufacturers to design pressure transmitters to meet practical requirements. This is especially essential for demanding applications.

The general conditions for the development of offshore oil fields are extremely difficult. Far from the mainland and at great depths, pressure transmitters are exposed here to high loads. Functional failure is extremely costly, since in the event of failure, the module has to be retrieved from the deep sea and then also reinstalled. It is essential to make reliable predictions in advance about unit functionality under the conditions to be anticipated. For this reason, the individual components of the pressure transmitter are first exposed to a mechanical simulation of the environmental conditions found in the deep sea.

Figure 1: FEM simulation of a sensor housing

The finite element method (FEM) is used in mechanical simulation. This is a common numerical process for examining the strength of bodies with a geometrically complex shape. The solid body to be examined, such as the housing of a pressure transmitter, is divided into finite elements, or partial bodies. This is therefore a physical modeling using computationally intensive software to determine whether the finite elements, and ultimately the overall structure also, would withstand those forces to be expected. Oil exploration is primarily distinguished by very high pressures. At a sea depth of 2,500 meters – by no means unusual in this field of application – a pressure of 250 bar is exerted upon the housing. In addition to this external pressure, the process pressure itself must also be taken into account, which can even be considerably higher (when pressure peaks occur, for example).

In the finite element method, therefore, no finished pressure transmitters are examined for their strength, but instead a modeling is performed as realistically as possible. If a solution is found that meets the specifications of the user, the product would then be tested out in an actual experiment, which will no longer be taking place virtually. In an individual pressure measurement solution for users in offshore oil production, this experiment in the pressure chamber is of primary importance. These hyperbaric tests validate the results of the finite element method and determine the load limit of the components or of the entire system. This ultimately ensures that users with special sensor requirements receive a product that performs reliably.

Figure 2: Micrographs of two sensor housings. Left: no pressurization. Right: after a hyperbaric test at 1,500 bar. No changes seen, the housing is stable.

Figure 2 shows the micrographs of two identical sensor housings. The housing shown on the left was not pressurized, whereas the right one was subjected to a pressure of 1,500 bar. This corresponds to a water column of 15 kilometers and thus much more than at the deepest point of the oceans. By optimizing the component using the finite element method, it can be modeled to withstand this enormous pressure. For comparison, the Mariana Trench is the deepest point of the oceans at 11 kilometers down. Pressure measurements taken even in the Mariana Trench itself should therefore pose no problems. The safety margin for most applications is thus very high and reliable operation is guaranteed.

Further applications of the finite element method

Mechanical simulations are not only useful for high-pressure applications. As already described elsewhere, temperature is an important influential factor in piezoresistive pressure measurement. Let us now take the exhaust pipe of a motor vehicle as an example. The temperatures here are very high and can exceed the limits of a pressure transmitter. In this application case, the finite element method would be used to investigate how the pressure transmitter must be designed so that no more than 150 °C of heat acts upon the measuring cell.

Mechanical simulations can also be useful in the low-pressure range. Mechanical changes, after all, have a much greater impact at low pressures. While measurement deviations in the mbar range are unlikely to be decisive in a high-pressure application, this is already a significant value for a measuring range below one bar. As an example, the connecting element between the measuring chip and the housing is usually an adhesive. If the torque is too high when mounting the pressure transmitter, this connection could be loosened or even just slightly altered and distortions would then be transferred to the measuring cell. This alone can lead to serious measurement errors. The properties of the adhesive used can also be modeled using the finite element method. The aim here, of course, must not be to find out the load limit of the connecting element and convey this to the user, but instead to find a solution that can easily withstand all possible torques applied during mounting.

The effort of mechanical simulations does pay off in the long run. Not only can products be designed to meet the required specifications, but this also allows for optimizing the design so that the products are as user-friendly as possible.

Position can influence the accuracy of pressure transmitters

Position can influence the accuracy of pressure transmitters

The accuracy of a pressure measurement can definitely be influenced by the position of the pressure transmitter. Particular attention to this should be paid in the low-pressure range.

When it comes to position dependence, inaccuracies can occur if the position of the pressure transmitter differs in practice from that used during the calibration process at the manufacturer. At STS, the norm is for pressure transmitters to be calibrated in a vertical position pointing downwards (see accompanying image above). If users now mount one of these calibrated pressure sensors in the opposite position, i.e. pointing vertically upwards, then inaccuracies may occur during the pressure measurement.

The reason for this is simple. In the latter position, the actual weight of the pressure transmitter will influence its precision. The membrane, filling body and transmission fluid act upon the actual sensor chip due to the gravitational force of the earth. This behavior is common to all piezoresistive pressure sensors, but it is only of importance in the low-pressure range.

Installation of pressure transmitters: Caution in the lower pressure ranges

The lower the pressure to be measured, the higher in this case the measurement error will be. With a 100 mbar sensor, the measurement error amounts to one percent. The higher the measuring range, the lower the effect becomes. Starting from a pressure of 1 bar, this error becomes practically negligible.

This measurement inaccuracy can be easily detected by users, especially when a relative pressure sensor is used. If users are working in the low-pressure range and it is not possible to mount the measuring instrument in the position in which it was factory calibrated, it should then be recalibrated in its actual position. Alternatively, users can also compensate for the measurement error themselves numerically on the control unit.

This additional effort can, of course, be easily avoided through competent application advice. Although STS pressure transmitters are calibrated vertically downwards as standard, it is easily possible to carry out the calibration in a different position. Our advice is to communicate the mounting position of your pressure transmitter with us in advance and you will then receive a measuring instrument perfectly matched to your application.

We will be only too happy to advise you!