Encoders in motor drive applications
What is the purpose of encoder? When is it needed? What types of encoders are being used? This post shall answer some of these questions.
VFDs are used to control the motor speed or motor torque. The actual speed is either estimated (speed sensorless control), measured with encoder or not used at all (simple scalar control). Besides speed signal the rotor position is required in certain applications.
Open loop versus closed loop speed control
Simplest scalar control (often called V/f control) does not have speed feedback. We might call it open loop control (feed forward). It just varies the stator frequency of electric motor and the motor is supposed to follow that frequency. We can also imagine it as direct on-line operation on a grid with variable frequency. Such scalar control does not allow very precise speed control as asynchronous motor always has certain slip that is load dependent. Synchronous machines do not have any slip in normal operation, but for those machines simple scalar control is anyway not suitable. Simple scalar control is sufficient for applications with low dynamics and lower accuracy on speed reference.
Closed loop speed control systems, such as direct torque control, vector control, advanced scalar control and other techniques, require information about actual motor speed. The actual speed is compared with reference speed (setpoint) inside the speed controller and a torque reference is computed accordingly.
How does the VFD get the information about actual motor speed? There are principally two ways:
– Speed estimation (observer inside VFD control)
– Speed measurement (encoder on motor shaft)
A) Speed estimation
Most VFDs use encoderless (sometimes also called more generally ‘sensorless’) speed control. What it means is that the rotor speed (and rotor position if required for control) is calculated (estimated) from the mathematical model of motor inside VFD control. This motor model is computed in real time with sampling time in the range of microseconds. A speed sensor is therefore not needed. It simplifies the interfaces and is considered as a robust solution. There is no need to install encoder on motor shaft and pull additional wires from motor to the VFD. However, the computational algorithms have their limits. At very low speeds the estimation is generally less accurate. It is usually sufficient for a short acceleration throughout this range. However, if the VFD shall operate at low speed continuously then a speed encoder is usually recommended by manufacturer.
B) Speed measurement with encoder
As described above, encoderless control is usually preferred in general purpose applications. However, some applications require speed encoder. Typical cases belonging to this category are:
– very high dynamic drives in wide speed range
– drives requiring high speed accuracy
– applications with high starting torque at standstill or very low speed
– fast flying start (VFD to catch a spinning motor)
The VFD manufacturer shall advise if speed encoder is needed for intended application or not.
Besides speed encoder some applications may also need a position encoder: These cases are mainly synchronous motor drives with high dynamics and precise control or synchronous machines with low saliency of rotor where the encoderless determination of rotor position becomes challenging.
Different terms are used with regards to encoders. We try to clarify few things.
>> Speed encoder (incremental)
Speed encoder is typically realized as incremental encoder, also called pulse encoder. It counts the pulses, most commonly on optical principle, and calculates the speed out of it. There are many speed encoders on the market. Important is the compatibility with the VFD (format of output signal) and resolution (number of pulses per revolution). The latter one shall consider the speed range of the motor. Typically incremental encoders with 10 to 16 bits are used.
The VFD control may use two different algorithms to process the signal from incremental encoder. One is suitable for low speed range, the other one for high speed range. The target is to maximize the accuracy throughout the speed range.
Figure 1: Incremental encoder – heavy duty design (courtesy of EMETA [1])
>> Position encoder (absolute)
Position encoder includes coding to determine absolute rotor position. Therefore also the name absolute encoder. Very common way to code the information about rotor position is Gray code (reflected binary code). Therefore, some people might call the position encoder a gray encoder.
Depending on the device a position encoder can also provide a speed signal derived from position so you don’t necessarily need two different encoders on the shaft if you need both speed and position signals and the provided accuracy is sufficient.
Physical principle of encoder
The encoders can be realized on several principles. Most common is optical encoder having a disk of glass or plastic that has transparent and nontransparent sections, a light source and a photo detection array (see picture below right). Other principles are magnetic absolute encoders (robust against dust and vibration), capacitive absolute encoders etc.
Rotary 3-bit absolute (position) encoder
Source: en.wikipedia.org
Absolute rotation encoder using Gray code pattern and optical sensor
Source: www.researchgate.net
Another principle of speed measurement, widely used in the past, is a tacho(generator). This is basically a dynamo giving output voltage proportional to rotational speed. The output signal is therefore analog and can either by used directly or converted into digital signal with A/D converter.
Take away:
– Primitive scalar control uses open loop principle and speed signal is therefore not needed. Scalar control is mainly used for asynchronous motors and square torque loads with low dynamics.
– Vector control, direct torque control and other advanced motor control principles need speed feedback signal.
– Speed signal is either computed in real time by motor model inside VFD control or measured with speed encoder mounted on a motor shaft.
– Quadratic torque applications usually operate as encoderless, i.e. the speed is calculated by the mathematical motor model from electric signals. Encoderless applications are robust and inherently immune against dust, EMC disturbances etc. They are the preferred choice for less demanding applications.
– Speed encoder is recommended for high dynamic applications and applications with high starting torque .
– Synchronous motor drives may also require a position encoder, e.g. if initial back spin shall be avoided.
– Speed encoders are also called incremental encoders or pulse encoders.
– Position encoders are also called absolute encoders or gray encoders.
– When selecting an encoder parameters such as speed range, resolution, accuracy, interface and compatibility with VFD, ambient conditions and installation shall always be considered.
<<Encoders in motor drive applications>>
References
[1] EMETA – Encoders “Made in Sweden”, https://www.emeta.se/?lang=en
