tel: +44 (0) 1223 413500 
info@cambridgeic.com 

News from CambridgeIC

Position sensing for motor control - what you need to know in three graphs

There are many sensing solutions for motor control; optical and magnetic encoders, Hall sensors, resolvers. Between them they provide a wide range of options with different performance in speed of response, design efficiency, durability and cost effectiveness. However all of the widely used sensors for motor control have strengths and weaknesses; it is important to understand these when selecting a sensor.

What to consider in selecting sensors for motor control

Whatever type of motor you’re controlling (stepper motor, servomotor, brushless/brushed) or whatever measurement point you are using for feedback, there are some basic trade offs to think about in selecting position sensors. Some of these are quite obvious: for example sensor accuracy, unit cost and speed of response. Other factors aren’t always obvious until further into a design. To look at the more complex questions in selecting sensors we’ve boiled them down into three graphs and three posts - if you don’t recognise your application in these scenarios, please get in touch!

Position sensors: how geometries affect integration costs.

In this first post, we’re going to deal with the relationship between design geometries and integration costs.

There are options for measuring every type of movement; either directly in the motor or by looking at an output i.e. position of a gear. At its simplest, linear position simply indicates where a target has got to along its track, while rotary position provides a measurement of how much a shaft is rotating.

targetapplication w500

Rotary sensing is done in two ways - either by mounting a target on the end of the shaft (known as ‘on-axis rotary’) or by mounting a target and sensor around the shaft (known as ‘through hole’ rotary)

EndAndThroughShaftCompare with labels

Finally, an arc sensor measures an angle less than 360° with both target and sensor positioned away from the rotation axis.

 

 CameraElevationApplication2017 06 16A

The costs of position sensing

For each geometry, if you know what you want to measure and the technical requirements you need for feedback (accuracy, speed etc) it’s tempting to believe that it is then just a matter of selecting the sensing solution that has the lowest unit cost. Unfortunately there are often additional integration costs that need to be taken into account and these can vary with the different geometries you might need. At a glance, the situation looks like this:

Position sensor selection graph 1

The reason for this is that many classes of sensor need very careful alignment, for example optical encoders and Hall Effect encoders. They do not tolerate large gaps or misalignment, and as a result they are often supplied as a packaged device with their own bearings which drives costs up and adds bulk. Even for precision solutions such as resolvers and LVDTs heavy expensive packaging is the norm. Some - especially optical sensors - will also need housing and seals to protect the sensor against long term damage from a dusty or dirty environment. Your product will usually have its own bearings, and to avoid a “fight” between the product’s bearings and the sensor’s bearings you’ll normally add a flexible coupling too.

These extra integration costs have even more impact in linear position sensing and arc position sensors, partly because linear and arc sensors are physically larger due to the greater measuring range and also because supporting a full range of linear sensors creates more product variants and more costs for manufacturers.

Integrating position sensing at low cost

By implementing precision coils into a PCB, CambridgeIC can produce resonant inductive position sensors that are highly tolerant of misalignment and dirty/harsh environments, without extra bearings, housing, packaging or couplings. The incremental cost for sensing arc position and for linear sensing is relatively low, because the only additional cost is a greater area of PCB, which is a relatively low-cost component compared to patterned magnets and optical scale.

The result is a sensing solution with no hidden extras and a total cost, especially where several axes are being measured, that is lower for high volume products.

So much for the impact of geometries on integration costs. In part 2 we will deal with how different sensor solutions deal with different environmental conditions.

 

If you find this useful but don't want to keep checking back, sign up for our regular round up. Sent out quarterly, it involves latest product news, along with hints and tips for engineers working with CambridgeIC products.

 

Get parts 2 & 3 with our quarterly roundup.

* indicates required