High-precision encoders integrated with bearings are large and bulky. For this reason, most industrial manufacturers use kit encoders that are assembled on site by their own technicians.
Aligning the rotary target with the centre of the shaft with a precision of several micrometres is difficult. But, it is the only way to reduce a significant sinusoidal angular error that is proportional to the remaining eccentricity value.
Metal target technology is usually only used for motor encoders. The rotor must move for self-calibration, so encoders do not provide full accuracy at power-up. Such encoders should not be used after a reduction gearbox as there is less movement at this point.
Our Solution
Our encoders are constructed using a target PCB with a precise pattern and a stator PCB containing an excitation coil and quadrature receiver coils. An electronics PCB is mounted on the top of the stator.
The advanced technology used by Cambridge Encoders reduces the need for alignment requirementsby more than a hundred times. The absolute kit encoders can tolerate an axial offset of up to 150 µm, making them easy to install.
Our self-calibration algorithm does not require rotation of the metal target. Our encoders provide you with perfect accuracy immediately after switching on. Self-calibration allows CamEncoders to achieve ten times better accuracy; 16-bit accuracy is achieved out of the box.
The developers of mechatronic systems are aware that their products' quality depends mainly on the quality of the encoders integrated into the rotating joints.
Inductive rotary sensors measure signals along the entire circular path and therefore offer a degree of symmetry that optical encoders lack.
The optical sensor:
A target misalignment of 15 µm can cause an error of up to 250″. It is possible to compensate for this error by using several distributed sensors and combining their readings as shown in the figure.
The inductive sensor:
The rotational symmetry in the multi-period sin/cos channel system of the inductive sensor makes it immune to alignment errors of up to 150 µm, keeping the overall accuracy of the system at 20″. This is a far better technique for suppressing alignment errors than the distributed optical sensor approach. Consequently, the use of inductive encoders greatly simplifies the installation of sensors in mechatronic systems and helps achieve much higher accuracy.
Precise micron-level alignment of the optical scale to the centre of rotation is achieved by using a custom-built adjustment rig for optical encoders. However, in actual mechatronic systems, your rotor is used with deep groove ball bearings that have a standard radial clearance of 15 µm. This clearance is required to compensate for interference fittings and thermal expansion. Non-zero radial clearance makes claims of accuracy that can be achieved with optical or magnetic sensors unfeasible in most real-world applications. Compare this to the 150 µm required by an inductive encoder and you can see the value of inductive sensing technology!
A sinusoidal error Δϕ caused by the eccentricity e in the measurements made by an optical encoder, as explained in the Heidenhain tutorial.
M= centre of graduation; ϕ = “true” angle; ϕ’= sampled angle.
Measurement error Δϕ as a function of the true angle ϕ. Eccentricity value e=15 µm, graduation diameter of optical encoder D= 24.85 mm .
Motor encoders
System engineers need to understand why certain encoder technologies are only suitable for motor control applications.
The sensor technologies used in encoders usually provide a pair of so-called sine/cosine channels. The ratiometric answer, ϕ= atan(Vsin/Vcos) , can be calculated from the voltage amplitudes measured in quadrature channels. Ratiometric sensors can be represented graphically by plotting the voltage in the cosine channel Vcos against the voltage in the sine channel Vsin.
Most ratiometric sensors are not error-free. As shown in the figure, the centre of the cosine/ sine curve can deviate significantly from zero. The deviation of the centre of the curve from zero implies non-zero breakthrough signals in both channels that are not related to the target rotation.
This significant error channel is typical of capacitive, Hall effect, and metal target inductive sensors. Resonant target technology shows a negligible shift from zero for the cosine versus sine curve’s centre. It makes a resonant target technology an exception and enables it to deliver much more accurate encoders.
All ASICs have implemented an algorithm that dynamically calibrates the breakthrough signal shown in the figure. They can also determine the difference in the gain of the sine and cosine channels to produce an accurate circular Vcos vs Vsin curve. Such calibration requires several revolutions of the target. The encoder output will have a significant linearity error until self-calibration is finished after power-up.
For many applications, self-calibration during target rotation is a good option. Such self-calibrated sensors have already been used successfully in motor control. Cambridge Encoders has developed proprietary solutions that extend the use of inductive sensors to other angle control applications.
A sinusoidal error Δϕ due to the breakthrough in the measurements of Vsin and Vcos. ϕ = “true” angle; ϕ’= sampled angle. The shift of the received signals from zero, when represented graphically, is similar to the eccentricity e described for optical sensors. However, it is a completely different problem.
Metal Target
Cambridge Encoders offer subcontracted services to develop sensors using Renesas’ IPS2200 ASIC analogue front-end. The cutomers of Cambridge Encoders receive object code for the microcontroller so they can make their own sensors without any further input from us. We can always help you change the coil geometry if you need to.
An eddy current principle: Rotary Target has a metal pattern that shields inductive AC magnetic field generated by an excitation coil.
CamEncoders offers a non-exclusive licence for the sensor design and the microcontroller object code.
Disadvantages:
Direct breakthrough is in phase with sensor signals.
Four times lower resolution.
Target rotation is required for sensor self-calibration.
In inductive sensors, the mutual inductance between the excitation coils and the receiving coils is close to zero. However, zero is an elusive number; non-zero breakthrough is one of the main error source. When target is rotated, the breakthrough signals can be measured as the deviation of the centre of the Vcos vs Vsincurve from zero.
In Cambridge Encoders solution, the last valid value of the calibrated breakthrough signals is read from the non-volatile memory at the start-up. The gain mismatch between the measurement channels is corrected during start-up using a method that does not require the target to be moved. Full accuracy is therefore available immediately after switch-on. As the target rotates, the shape of the receive signals is used to continuously update the gain mismatch and calibrate and write to the memory the updated value for the breakthrough signals.
Cambridge Encoders offer kit encoders that consist of several PCBs and are installed in the mechatronic system on site. Our products include bearingless encoders and ruggedised encoders with integrated bearings. Customers can choose their form factor and we can develop customised solutions for them, for which we charge development fees.
A resonant circuit consists of a precisely shaped coil and a capacitor. It is tuned to resonate at the frequency of excitation coil drive.
Advantages:
The resonant coil signals are 90 degrees out of phase, so they are in quadrature with the direct breakdown signals.
Higher resolution and accuracy compared to metal targets.
Target does not need to be moved for sensor self-calibration.
Disadvantages:
Greater installation height
Commercial ASICs are not available; the analogue design is implemented in discrete, commercially available components.
Cambridge-based encoder companies have perfected resonant target technology; German encoder companies use metal targets. The resonant target technology has a unique advantage – it reduces unwanted breakthrough from excitation coil to receiving coils by a factor of about fifty, making it an insignificant source of linearity error.
With our inductive sensor electronics, we continuously check and calibrate the gain mismatch between the measurement channels, regardless of whether the target is stationary or rotating rapidly. The receive channels are always self-calibrated, regardless of the target’s range of motion. Encoder applications benefit from this approach especially when the target movement is severely restricted most of the time. Resonant target technology is recommended by Cambridge Encoders for precise measurements following reduction gearboxes.