The Progression to Low Cost Current Sensing
This article highlights Pewatron on cost reduction opportunities in measuring the traction current in universally used three phase motors.
In most industries there is continuous pressure to pull cost. This is particularly true for the automotive environment. Environmental pressure means that hybrid or pure electric vehicles will be sold in escalating numbers. Higher production volumes will mean considerable design refinement in all areas of the drive train.
This article focuses on cost reduction opportunities in measuring the traction current in universally used three phase motors. Currents are generally up to a peak of 2000A and battery voltages two or three hundred volts. Phase current measurement is required as an essential part of the variable speed drive algorithm. In order to optimise efficiency, phase voltage, current and phase angle measurements are required. Additionally current measurement is needed to protect against overload and check for system faults. In simple terms with a three phase motor, just two current sensors are required as the current in the third phase can be derived from the sum of the other two phases. However, when system safety and reliability are concerned, it is beneficial to use three current sensors so that system cross-checks can be performed.
For a number of reasons it is beneficial to measure actual phase current rather than attempt to derive it from bridge leg current. No great precision is required. Motors are not precise devices so a few percentage errors have little effect on system efficiency on smoothness of drive or perhaps noise. On the other hand, accuracy does need to be maintained over all speed and loading ranges and the four operational quadrants.
The Conventional Approach
This approach uses three separate current sensors which have the necessary precision over the required current and temperature range. They probably also need a frequency response greater than about 30kHz to “see” the PWM oscillations and respond quickly to allow short circuit protection.
Figure 1: Conventional 3 Phase current sensing – 100%
These sensors will probably be open-loop hall devices with a magnetic core. They should be reasonably tolerant to conductor position within the sensing aperture. The 3 sensors could cost about $ 120- 150.
Coreless current sensors The next generation of sensors for the traction environment will be coreless. An array of magnetic field sensors is positioned around the aperture replacing the core. This reduces size and weight. (Size and weight are important qualities for EVs.)
Figure 2: Coreless 3 Phase current sensing - ~ 66%
Typically, coreless sensors have excellent linearity as there is no core saturation effects. Also their frequency response can be over 100kHz.
Three Phase Sensor
The next cost reduction progression is to go to a three phase sensor. With coreless this is particularly attractive. It also saves 2 connectors as just one 3 phase connector is needed. The sensors can be laid out on a single PCB. If the environment allows, the housing may not be necessary. A thick conformal coating would offer protection against humidity. Probably electrostatic screening will be required in order to improve the signal to noise ratios.
Typically, coreless sensors have excellent linearity as there is no core saturation effects. Also their frequency response can be over 100kHz. Three phase sensor The next cost reduction progression is to go to a three phase sensor. With coreless this is particularly attractive. It also saves 2 connectors as just one 3 phase connector is needed. The sensors can be laid out on a single PCB. If the environment allows, the housing may not be necessary. A thick conformal coating would offer protection against humidity. Probably electrostatic screening will be required in order to improve the signal to noise ratios.
Figure 3: Three Phase coreless PCB with screening ~ 20%
The cost now plummets down to about less than a third. The assembly can be very compact – perhaps fully integrated into the speed controller.
Cost Reduced Three Phase Sensor
The reason behind the magnetic field sensor array is to provide good immunity to stray external magnetic fields. It also makes the sensor largely immune to conductor position. Now, if conductor position was fixed and there were no other nearby conductors other than the motor feeds, an array of sensors may no longer be needed. A single magnetic field sensor for each phase could well be sufficient. There will be cross-talk between phases but this will be systematic so can be compensated with a suitable algorithm.
Figure 4: Three Phase coreless PCB with screening and compensation ~ 15%
This option could pull a further significant percentage out of the cost. The negative with this option is that a fairly sophisticated calibration process could be necessary. However, once the algorithm was fully developed, the calibration process would be fully automated. The end result is that the hardware associated with current sensing becomes trivial.
Like many simple systems, careful design is required to assure necessary performance. Raztec has many years of current sensing design skills and their design-in partner Pewatron with experienced specialists are an excellent position to consult for particular applications. The supply partners offer a wide range of magnetic field sensors to choose from and are familiar with critical supply aspects demanded by the automotive industry or other similar application fields. Customer specific solutions are therefore very much welcome and the earlier the engineers are involved, the better the solution can be tailored to the needs.
The progressions documented in the article hints at some disruption for traditional current sensor suppliers. The progression listed will happen. It is more a matter of “when” rather than “if”. Raztec and Pewatron can help customers smooth and accelerate the “when” process.
This article originally appeared in Bodo’s Power Systems magazine.
About the Author
Warren Pettigrew holds a Bachelor's Degree in Electrical and Electronics Engineering at Canterbury University. He worked at Raztec as the Chief Technical Officer. Currently, he works at the Dynamic Controls as the Senior Research Engineer.