What’s best: Standard or custom?

The first question that needs answering is why you are measuring something in the first place? This may seem rather obvious, but it is vital that you understand just what it is you are measuring and why it is being measured. This needs to be clearly defined right at the outset of the design process. But, no matter what it is you want to measure – displacement, pressure, temperature or whatever – the principles for sensor selection are generally the same. "Once you have defined what it is you want to measure, you then need to understand what environment the sensor will operate within," says John Tyrrell, managing director of Ixthus. Ixthus, the bespoke sensor manufacturing arm of Heason Technologies and works closely with Variohm, which distributes standard sensors. "The core questions then are, does it need to work in a harsh environment and are there any space constraints? And, of course, you need to be thinking about cost," he adds. And it is all too common to get it wrong, as Ixthus discovered when a customer came in and brought a custom designed force sensor: effectively a stud with a M8 thread on both ends, strain gauges in the centre. "They were using it to measure compression loads up to about 5kN," says Tyrrell. "It worked, but they wanted an improvement. So we supplied them a COTS load cell that was 60% of the original purchase cost that worked three times as well. "Sometimes engineers come to us with a pre conception and idea and that is the only way it will be solved." While bespoke sensors can be made to specific requirements, they are generally more expensive. If you can compromise on some parameters such as resolution or accuracy, then it will save you time and money. However, if there is a strong driving force such as the need to measure in a harsh environment, tight space or to exceptionally high resolution, then it will usually lead to a customised design. Bespoke sensors should also be considered for higher volume applications to take cost out of components and get a lower unit cost if many sensors are likely to be purchased on a regular basis, such as in the automotive industry. Once you are comfortable that you know what you want the sensor to measure, in what environment and to what accuracy, then it is time to look around for a competent and unbiased view to the sensor technologies available. "If you start with a sensor in mind, these tend to be self fulfilling and you can build up a specification around a technique or technology," says Chris Jones, managing director of Micro Epsilon UK which supplies both commercial off the shelf and bespoke sensors. "As a result, sensor suppliers are increasingly taking on a consultancy approach."If you are not comfortable inviting in a sales consultant or supplier at that stage and you want to do some web research, a good method is to look at comparative technology journals which explain what the strengths and limitations of each technology are – so you have a bit of a flavour of what they do. Complete customised solutions are usually expensive. And in many cases a standard solution can be modified or adapted for an application. It is also possible to get more out of the sensor using clever software, a different calibration, or mechanical mounting. For exmaple, if you want to get better accuracy you can use improved electronics or manufacture the sensor from better components. Jones advises: "We can usually tweak and improve certain parameters relatively easily. Much of it is about understanding the limitations of each technology. "It is often possible to optimise a sensor to reduce measurement error. Often, the biggest error comes from temperature errors which drift the electronics and cause thermal expansion in the actual material of the sensor. "You can characterise a sensor by putting it in an oven and looking at how output changes with temperature," says Jones. "The end customer can get the sensor and calibrate it over different positions to work out linearity error and correct for this. With a PC or smart PLC you can create a 'lookup table' to correct for those errors by introducing a temperature sensor to the system." Micro Epsilon re-engineered capacitive sensors for Attocube, a supplier of nano-positioning systems for Atomic Force Microscopy, to measure the travel displacement of test items that need to be moved in the nanometre range, enabling surface topography data to be obtained. Although still a COTS sensor, Micro Epsilon has been able to get the error from 200 to just 12 parts per million (ppm) per °C. The non-contact capacitive displacement sensors measure the travel distance of test items down to nanometre resolution. The sensors also have to operate in extreme conditions, with ambient temperatures of -270°C and in an ultra-high vacuum. Because the sensors are expected to perform just as well at -269°C as they are at room temperature, Micro-Epsilon used special materials for the sensor and the cable, which due to its very low thermal expansion properties, provide very stable sensor measurements. Mounting a sensor correctly can be a bit of a black art. If a sensor is not mounted correctly and it is not in the direct path of what it is measuring, then everything on your readings is going to be inaccurate. Additionally, it must be on a rigid support so it does not vibrate during measurment. Trevor Astin, senior sales manager of ASM UK says: "The mounting and location of sensors is an extremely important factor in getting the most from your sensor. For example, if you take a traditional rotary potentiometer with an output shaft, it is extremely critical that the shaft is correctly aligned in all axes to prevent damage to bearings or the potentiometer track. A misalignment can cause the wiper to lift from the track which in turn will cause the signal to become open circuit."