Top 10 Position Sensor Mistakes

With so many position sensors available these days, it is of little surprise that engineers may occasionally make a mistake when it comes to choosing the position sensor for their application. Mark Howard, Managing Director at Zettlex, suggests the top 10 position sensor mistakes.

• Using contact sensors where non-contact sensors should be used

Potentiometers are still a common position sensor and there are some applications where they work well.  Applications where they don’t work well are ones with prolonged vibration or foreign matter – especially fine sand, dust or water. Vibration or foreign matter quickly degrade the potentiometer’s track and contacts.

You can learn more about potentiometers here – Why do engineers dislike potentiometers?

• Optical or capacitive sensors in wet or dirty conditions

Optical and capacitive sensors can be unreliable in dirty conditions. High resolution devices are particularly vulnerable and beware the phrase “….but it’s a clean, sealed environment”.  One word – condensation. One more word – seals sometimes fail (OK, 3 words).  If you’re lucky, a dirty environment will cause an optical or capacitive sensor to stop working.  If you’re unlucky, it’ll carry on working but giving the wrong signal. If you’re very unlucky, the consequences can be catastrophic.

Useful article if you would like to understand the different types of position sensors available – Choosing the right Position Sensor

• Not understanding the difference between resolution, repeatability, linearity and accuracy

Come on admit it, you’re not sure.  The majority of engineers seem to have skipped college on the day this was covered. The full technical definitions deserve a paper by themselves but a quick way to illustrate the point is with an angle encoder with say 1000 counts per rev. It doesn’t mean it’s accurate or linear to 1/1000th of a rev. It just means its output signal changes 1000 times per rev. That’s all.

Want to know more? Here is a technical article providing more detail – Accuracy, Resolution, Repeatability and All that Stuff

• Not reading the small print of the sensor’s spec

The headline might say ‘accurate to 0.01 degrees’ but buried in the datasheet’s small print is ‘when concentricity of sensor rotor is <1micron and read head installed to a position tolerance of +/1micron and temperature is 21.00oC’.  This is particularly true of optical encoder kits where the designer is responsible for mechanical packaging.

• Not designing for connectors and cables

Many design engineers will arrange the main sensor body and its mechanical fixings but forget about cable routings and connectors.  Cables and connectors are a crucial element of any position sensor – don’t leave them to last when you’re planning the layout.

• Confusing incremental and absolute sensors

If you’re not sure what the difference is then the acid test is if the position sensor needs to carry out a calibration step at power up, it’s an incremental sensor. Beware of sensors that claim to be absolute but actually need a small amount of motion to produce an absolute output – they’re incremental but with clever marketing.  Absolute sensors output correct position right from power up.

Here is an article discussing the differences in more depth – Incremental versus Absolute

• Measuring position indirectly

If you are measuring the angle of a shaft and the sensor is mounted on the shaft that’s direct measurement.  If the sensor is arranged at the end of a gear train and shaft position is being inferred from the position of a gear, then that’s indirect measurement.  As a general rule, high accuracy indirect measurement is just not practical. When examined in detail there are so many intermediate effects such as backlash, differential expansion etc. that the fidelity of the measurement is degraded.

• Electrostatic build up

Capacitive encoders are susceptible to the build-up of electrostatic charge due to motion, rubbing or rolling. Plastic or even steel bearings can generate static. That means shafts and bearing assemblies may need to be earthed.  It’s another small print thing.  Watch out – earthing a rotating shaft can be difficult and expensive.

• Magnetic noise and hysteresis

It may be obvious but magnetic position sensors measure position by measuring magnetic field strength.  Electric cables, motors, transformers and magnetic brakes also generate magnetic fields. If they are close by they will generate noise that may be picked up by the sensor.  Also be careful with the stated resolution of magnetic sensors.  There are instances of resolutions of 14bits but repeatability of <10bits – this is because of the inherent hysteresis (tantamount to backlash) of any magnetic field.

Further reading on magnetic sensors – Magnetic Sensors with Inductive

• Using a slow sensor in a dynamic system

There’s a term in a position sensor’s datasheet that’s often overlooked:- update rate. In other words, how quickly the sensor’s measurements are carried out and reported.  Many sensors have an update rate of <100Hz.  In a dynamic environment, where the position of a rapidly moving target is to be measured, the update rate realistically needs to be >1kHz and preferably >10kHz, otherwise the dynamic error due to the time delay between actual position and signal will be large.

If you haven’t already be sure to connect with us on social – Twitter / LinkedIn / Google +