What's our Technology?

Our control technology learns about the motor and its use


Our motors operate without expensive and unreliable sensors taking up valuable room in the motor housing.

Low Torque Ripple

Our patented solution for low-speed control and eliminating noise allows your motor to operate smoothly without


Our software starts by learning the characteristics of the motor


Our software does away with the need for position sensors (encoders and resolvers) by calculating the rotor position from measurement of current and voltage.

  • Solving Kirchoff’s 2nd Law to give us the winding emf
  • Algorithmically calculating motor position from winding emf
  • Continuously updating the motor position 10,000 times a second.

At low speed, the problem becomes even harder, as there is no winding emf, so we:

  • inverter calculates the winding inductance
  • from the winding inductance we can calculate the position
  • continuously updated 10,000 times a second

The winding inductance is hard to measure when there is an emf present so it is only used for low speed.

Our technology offers us a number of advantages

  • Smaller. Our motors can be smaller than those with sensors as sensors are often bulky and add volume, wires and airgaps (for EMC complience) to the motor.
  • More reliable. Our control software is completely independent of temperature and voltage. As we do not use sensors, our control algorithms has to work out the temperature (and therefore the resistance and effective permeability) rather than measure it directly. The sensor cables can also hinder the lifetime of the product, eliminating these simplifies the product.
  • More cost effective. Motor encoders and resolvers can be up to 10% of the cost of the motor, we eliminate that cost using our sensorless control technology.

Low Torque Ripple

To achieve low torque ripple we start with the motor design link.

  • Sinusoidal current achieved in conjunction with our sensorless control ensures that there is no abrupt changes in electromagnetic force leading to reduced torque ripple.
  • We avoid overshooting torque by using fine tuned algorithms that only apply the current that it is needed, no more, no less.
  • Vibration cancellation by eliminating torque and radial force harmonics.

A motor with less torque ripple runs smoother than it would otherwise, this has a whole host of benefits, including:

  • Reduced acoustic noise. Torque ripple causes the motor to accelerate in uneven ways, causing mechanical resonances – which create acoustic noise unless eliminated.
  • More accurate speed and position control.

We are pioneering the use of deep reinforcement learning in motor control. This technology comes with three distinct advantages:

  • Smaller and cheaper processors. The size of the embedded control software is smaller as only the trained weights of the motor are uploaded to the chip.
  • Predictive torque control. With deep learning the motor control can be predicting events that a conventional controller would not have.
  • Zero stalling. Conventional controllers can sometimes disasterously fail if something unexpected happens. With deep learning the algorithm will take a guess rather than stall even if the state is soemthing it has not seen in training


Our sensorless control forces us to push the boundaries of adaptive control, by having no sensors we need to learn the characteristics of the motor during operation. The inverter is continuously monitoring the motor and the load.

  • Using cross-calculations of emf and inductance we are able to determine the motor design.

Adaptive control is one of the technologies we are most excited about, here’s why:

  • Our software can generalise to uses in different operating conditions.
  • Our software learns about the motor and adapts to its use. We can monitor the condition of the motor and notify the user when parts degrade.
  • We can output real-time data of the motor’s running to improve the process control.