- Register



10 dumb ways a position sensor can mess up your project

27 March 2017

Measuring position or speed accurately and reliably is a tricky business; selecting the wrong sensor or developing the wrong installation often results in overly complex design, poor performance, high costs or unreliability. Mark Howard, MD at Zettlex, lists the ten most common mistakes – so you can avoid them in your next design project

1. Misunderstanding the measurement performance

If you’re using a term like “pretty accurate” or “fairly precise” to select a position sensor then you probably haven’t nailed down the required performance. Get this basic aspect of your design wrong and you’ll either be paying too much for an over-specified sensor or you’ll never achieve the required performance......and you’ll have to start again.

There are three key parameters for position sensors – resolution, repeatability and linearity. Resolution is the smallest change in position that the sensor can measure reliably. Repeatability is the maximum difference in the sensor’s output when it moves away from one position and comes back to the same position. Linearity is usually tantamount to accuracy (ignoring any off-sets) and is the maximum difference between actual and measured position.

The greater the resolution, repeatability or linearity – the higher the sensor price. Sensor price is heavily related to linearity. For many engineering systems, it’s actually repeatability which is the key measurement parameter, not linearity.

Tip: specify measurement performance carefully.

2. Underestimating the cost of failure

All design engineers are under pressure to minimise cost. However, the key to success is to minimise overall costs – not just component costs. Whilst it’s tempting to install the cheapest possible sensor, any savings will quickly disappear if it causes failure in the field. This can lead to equipment down- time; lost production; disruption; cost of equipment dis-assembly; cost of technician travel and time; loss to reputation. Don’t under-estimate the last one if you’re an equipment manufacturer. As a rough ball-park, the replacement of a failed sensor in the field will cost at least 10 times and maybe as much as 1000 times the original cost of the sensor.

key to success is to minimise overall costs – not just component costs

Tip: Don’t skimp on quality – it’s a false economy.

3. Misunderstanding the safety requirements

Every position sensor has the potential to go wrong at some point during its life. Often, if a sensor gives out no signal or an error code then the effect might not be so bad. The result may be a minor inconvenience and the host control system might revert to a fail-safe state. However, if the failure mode is such that a sensor gives out a credible but incorrect signal (think aircraft aileron control) then equipment malfunction can be dangerous and even catastrophic.

As with cost of field failure, skimping on safety is never a good idea. That’s not to say that you should overspend on an over-specified system; the trick is to specify the right system – by matching the position sensor and control system with the probability and impact of failure.

At the simplest level, there may be no requirement for any additional design steps or backups. However, as the probability and impact of sensor failure increases then further mitigating steps are required such as sensor built in test; error handling; periodic inspections; external reasonableness tests; electrical redundancy and avoidance of common failure modes.

Tip: Think carefully about failure modes and mitigate the risks to safety.

4. Selecting a potentiometer that wears out rapidly

Despite the trend towards non-contact sensors, potentiometers remain the most common position sensor. They use electrical contacts sliding along a resistive track and they can work well in benign environments. They are subject to wear and short life if the contacts get dirty or there is protracted vibration at a fixed position. The life of most potentiometers are specified for a maximum number of cycles but if a system is vibrating at say 10Hz then its life might be over in a matter of days.

Tip: Pots + harsh vibration are rarely a good combination.

5. Optical encoders in dirty environments

Optical encoders are a common form of position sensor and can provide accurate and reliable results. Their level of precision can be staggering due to the tiny features that can be etched on to the optical gratings. These features can also become susceptible to failure of the optical path from dust, sand or other foreign matter. In the best case, an error code or ‘no-read’ will be generated. In worst-case, the resulting position signal is incorrect. Optical encoders are not well-sited to dirty environments, whereas magnetic or inductive encoders (incoders) are unaffected by most foreign matter.

Tip: Optical encoders are usually not well suited to dirty or wet conditions.

6. Underestimating installation tolerances

Position sensor datasheets are, of course, keen to boast high levels of accuracy. What they are less keen on promoting is the tight tolerances that might be required to achieve the stated measurement performance. Make sure you read the datasheet’s small print – especially those for optical ring encoders where a read-head may require installation to tolerances of <10microns to achieve the head-line stated accuracy. Don’t ignore the effect of tight installation tolerances on material or assembly costs.

Tip: Make sure that the tolerances for installation are achievable and cost effective.

7. Calibrating each individual sensor

A common technique to counteract lack of accuracy from a (low-cost) position sensor is to calibrate it against a higher accuracy, higher cost position sensor. The calibration produces a look-up table so the output from the low-cost sensor is corrected by the host control system. In some situations (e.g. high precision electro-optics or weapon pointing systems) this works well. These are sophisticated manufacturing and service environments with products that will only undergo highly skilled service or maintenance. If this technique is used, remember that as soon as the sensor is moved or replaced, the system will need to be re-calibrated. This approach rarely works in less sophisticated environments or where the position sensor is required to be replaced, moved or uninstalled during the life of the host equipment.

Tip: Calibrating individual sensors? Make sure your supporting operations are up to the job.

8. Inaccuracy from indirect measurement

A position which is inferred rather than measured directly, is referred to as an indirect measurement. One example might be measuring the angle of a shaft by measuring the angle of a motor shaft, which drives a gearbox, which drives a coupling, which drives the shaft. Indirect measurement is never as accurate direct measurement. A surprisingly large number of parameters come in to play whenever indirect measurement is used such as gear-backlash, misalignments, differential thermal expansion and so on. Whenever possible arrange the design to measure directly rather than indirectly.

Tip: Whenever possible measure position directly – don’t infer.

9. Capacitive encoders in wet environments

Capacitive encoders are based on the capacitance between a capacitor’s plates as they displace relative to each other. Snag is, capacitance also varies with moisture, temperate, electro-static build up and humidity. As with optical encoders, capacitive encoders work well in clinical conditions but dirt, dust, grease, condensation and static can produce incorrect sensor signals. Capacitive encoders are seldom a good choice for wet environments (condensation is a well-known problem), whereas magnetic or inductive encoders (incoders) are unaffected by most liquids.

Tip: Capacitive encoders can be unreliable in dirty or wet environments.

10. Forgetting about cables & connectors

To minimise failure, cables should be tightly secured, gently radiused and non-flexing

Statistics show that cables and connectors prove to be as great a source of field failure as position sensors themselves. To minimise failure, cables should be tightly secured, gently radiused and non-flexing – especially in harsh shock and vibration environments - so that connectors remain unstrained and conductors unbroken. In wet or dirty environments the number of connectors should be minimised through the use of integral cables. Cables should also be rooted away from sources of electromagnetic noise; extreme temperatures or the harshest environments. As with other factors, choosing the cheapest possible cable and connectors is often a false economy.

Tip: Make sure your design considers cables, cable routings & connectors.