One of our first main goals is to develop a fully working, sensor less motor drive application from scratch. This includes the necessary power electronics, measurement cards and controller.
Parts and equipment has been ordered, and it is expected that we are ready to start prototyping PCBs in January next year.
The motor drive will be based on SiC MOSFETs which offers extremely low losses and the possibility for high switching frequencies.
To provide an overview of such a process, the main steps and their challenges are listed below. Each item will receive a dedicated post to show what we have chosen and why.
- Dimensioning the electric motor:
This step is usually performed using classical physics to determine the maximum power required to handle the load given certain specifications. As this is a lab project, this item is really not applicable. Instead we are aiming to use a 1 kW induction motor in the initial phase.
- Decide rated current and voltage for the power electronics:
This is further given by step 1, but as this is a lab project, we decided to shoot for 20 A as the minimum current and 400 VAC RMS input voltage.
- Select power electronic components:
Three main items are required and all three have to meet the minimum requirements regarding voltage and current:
- Diode bridge for front-end AC-DC rectification,
- energy storage in the form of a capacitor and
- a transistor bridge to syntesize the output voltage and a gatedriver to activate each of the transistors.
- Selecting measurement components
Although we refer to it as a sensor less motor drive, sensors for voltage and output current is needed to understand the motor's position and by that knowledge, create a voltage which will cause it rotate a step further. The term sensorless actually relates to rotor position sensors, such as flux sensors, encoders or resolvers. This we are planning to do without. It is perfectly possible to determine the rotor's position using only voltage and current measurements.
- Creating circuitry to condition the voltage and current measurements
"OK, so you've got a voltage sensor - that don't impress me much", as Shania would have said if she carred for motor drives. The measured voltage (or current) often need to be conditioned (transformed) in order to be useful for the analog to digital converter. For the voltage, it is often not possible to for the voltage sensor to accurately measure several hundred volts directly, so we step it down using resistors. On the output side, the voltage sensor might send a voltage level which is outside the range of the A/D converter, so further scaling is necessary. This we normally fix using op-amps.
For this project, we have decided to create separate PCBs for current and measurement for the sake of being flexible. It makes it easier to update/change the voltage sensor without having to rebuild a new current sensor PCB as well.
- Programming the motor controller
"Finally!", you might think, "now we have all the parts we need - just insert a start button and the drive is ready". Well, not so much. Step 1-5 was the easy part. Actually controlling the motor is the difficult part. This has to be programmed in both high-level languages as C++ for the processor and assembly like languages such as HDL for the FPGA chip. Eirik, my co-worker at this site, has some experience from his master's thesis, but myself need to learn a lot during this project in order to fully understand the inner workings of a motor drive control system.