Let’s start with a quick review of how motors work.

Simply stated, motors consists of two parts: a stator (stationary part — usually contained within the outside shell of the motor); and a rotor (rotates on a set of bearings) and has a central shaft on the rotor that is accessible outside the housing for attachment of pulleys, pumps or other working equipment.

A simple stator has three sets of electromagnets, spaced 120° apart, which can be turned on and off. Usually this is done by ramping up the current from zero to a maximum, and then back down to zero again in a sinusoidal fashion. There are also magnets (or electromagnets) on the rotor portion of the motor that are designed to match the configuration of the stator magnets. As the sinusoidal currents go around the stator in order, they push the rotor magnets ahead of them causing a rotating torque.

Of course many variations of this concept are available, but this simplified description should serve for now. As an aside, typical industrial power is between 110 V and 240 V, usually at 60 cycles per second (Hz). This built-in 60 cycle power source is sinusoidal and can be used to drive the motor directly. Motors that use the line voltage directly are known as synchronous motors. The limitation of using this readily available source is that the resultant RPM of the motor must always be some multiple of the line frequency — 60 Hz — and the base speed is fixed.

Generally, all of the electronic control of switching currents from one set of magnetic poles to the next is accomplished by a device known as the motor drive. For large motors it is not hard to imagine that there is a lot of energy to be controlled, so the timing is critical.

In addition, as it has been described here, the control is analog. This means that a lot of energy is lost in the form of heat as the motor rotates. These two limitations: controlling the motor speed, and energy lost in operating the motor drove engineers to look for a digital solution. The result is known as a variable frequency drive or VFD. Being digital, it must have a DC voltage as an “on” state. This is how it is accomplished.

  • Step one: Convert line voltage so that instead of being 120 V sinusoidal, it becomes 0-120 volts ½ sine. This is done using a full wave rectifier.
  • Step two: Smooth out the rectified voltage using capacitors. The voltage is now approximately 0-120 V DC with a slight periodic ripple at the peak voltage.
  • Step three: Use a DC-to-DC converter to remove the ripple and now you have a DC voltage appropriate for use as a VFD.

This may seem like a lot of extra work but a DC voltage is needed to use IC components. This sort of AC to DC conversion is common and in reality all of the steps take place concurrently often in pre-packaged converters.

Steps to converting an AC sine wave into a fixed DC voltage.Steps to converting an AC sine wave into a fixed DC voltage.The reason for going through this work is that DC components can be switched on and off very quickly and efficiently. This allows for the VFD to be configured so that its electric motor can be used at a variety of speeds and torque values via simple programming steps. By using a technique know as pulse width modulation (PWM), the VFD signal can be turned on anywhere from 0% to 100% of the time. VFDs typically have a normal operating frequency between 1,000 to 20,000 Hz. This is the number of discrete signals (either on or off) that it can produce per second. By programming the on and off states to deliver the energy to the motor coils in an increasing amount, followed by a decreasing amount, it can simulate a sine wave drive to the motor as shown in the illustration below.

Now that the drive is no longer constrained to just 60 cycles per second. The speed and torque of the motor can now be matched better to the load, making the whole system more efficient.

The heart of a VFD is the switching device used to control the current and switching frequency. It is a device known as an IGBT (Insulated Gate Bipolar Transistor). It is used for many reasons.

  • High switching frequencies. An IGBT can switch on as quickly as 400 ns.
  • Low switching current. In the on state IGBTs have a low resistance so the overall losses to heat generation make it more desirable than the alternatives.
  • High current driving. Collector to emitter currents as high as 100 A are possible and if that isn’t enough, two or more IGBTs can be put in parallel to drive even more current.
  • Competing semi-conductor choices, BJT and FET, can’t match all of these capabilities

Using PWM signal modulation to simulate a partial sine wave. Using PWM signal modulation to simulate a partial sine wave. Despite all of the good things associated with using VFDs for motor drives, there is one area of caution. Delivering the energy via “pulses” is sort of like ringing a bell. The impulse causes a resonance every time there is a change from zero to the operating voltage. Due to the capacitance and induction present in the wire bundles of the motor, the voltage will overshoot, sometimes to voltages as high as tens of thousands of volts for a very short duration. Providing enough insulation on the wire used to build the motor can usually handle that.

However this high voltage spike can also appear on the rotor, due to inductance. If the motor bearings are not properly isolated, then these short, high volage spikes can cause micro-pitting in the bearing races. Over time, this can cause a premature failure of the bearings. This is a known phenomena that motor manufacturers are aware of, so they design accordingly. It is possible that during the installation, the end user may not be as informed about this potential failure mode and inadvertently provide a path to ground through the bearings.

Consequently it is very important to understand and follow the manufacturers instructions when it come to installation and good grounding practices. VFDs have become so reliable that they have pretty much become the standard drive for large industrial motors today.