EC-12.4 Power Devices and Switching

What this is and why it exists

The device physics module already covers the diode and the MOSFET, as rectifying and amplifying devices.

This topic asks a different question of the same parts. What happens when they carry tens of amperes and are switched on and off thousands of times a second?

The answers, switching loss and thermal limits, are what power electronics is actually about. Nearly every design decision in the next two topics is a consequence of one of them.

The vocabulary

• Switch mode — operating a device fully on or fully off, never in between.
• Conduction loss — power lost in the on-state voltage drop.
• Switching loss — energy lost during each transition, multiplied by the rate.
• Thyristor — a device turned on by a gate pulse and turned off only by the circuit.
• Insulated gate bipolar transistor — a device combining voltage-driven control with a low on-state drop.
• Gate charge — the charge that must be moved to switch a device.
• Reverse recovery — the brief conduction of a diode after it is reversed.
• Thermal resistance — the temperature rise per watt across a thermal path.

The mental model

A power device is used as a switch and never as an amplifier, and the reason is arithmetic.

A device half on has a substantial voltage across it while carrying a substantial current, and the product is heat. Controlling a kilowatt that way means dissipating a large fraction of a kilowatt in a package the size of a matchbox, which is impossible.

Fully on, the voltage is small and the current is large. Fully off, the voltage is large and the current is nearly zero. In both states the product is small. Everything else in power electronics follows from the decision to live only in those two states.

The device family is short, and the four figures that decide the choice are always the same. How much voltage it blocks, how much current it carries, how fast it switches, and whether you can turn it off.

The diode conducts one way and is not controlled at all. The thyristor is turned on by a gate pulse and turned off only when the circuit takes its current away. That suits alternating supplies and rules it out of most switching converters. The power MOSFET is voltage driven and very fast, with an on-state that behaves like a resistance and so worsens at higher voltage ratings. The insulated gate bipolar transistor keeps the voltage-driven gate and adds a lower on-state drop at high voltage, at the cost of switching speed.

Losses split into two kinds, and keeping them separate is the whole of thermal design.

Conduction loss is the on-state drop times the current. It does not depend on how often you switch. Switching loss is the energy lost in each transition times the number of transitions per second, so it grows directly with frequency.

That single fact sets the operating frequency of every converter. Higher frequency means smaller inductors and capacitors, which is why designers want it. It also means more switching loss, which is what stops them.

Driving the gate turns out to be harder than the schematic suggests. Turning a large device on quickly means moving its gate charge quickly, and that means supplying amperes for a fraction of a microsecond.

A weak driver switches the device slowly, and slow switching means more time spent in the lossy region between the two states. A device running hot for no visible reason is very often a gate drive problem rather than a device problem.

Reverse recovery is the trap in the diode. A diode carrying forward current does not stop conducting the instant it is reversed. Charge must be removed first, and during that time it conducts backwards.

That recovery current flows through whatever device is switching on. It appears as extra loss in the transistor rather than in the diode, and as a current spike that radiates interference. It is a major reason for choosing fast recovery diodes and for the interest in materials where the effect is almost absent.

Thermal design closes the topic and it is one circuit. Heat flows from the junction to the case, from the case to the heat sink and from the sink to the air. Each step has a thermal resistance in degrees per watt.

Treat power as a current and temperature rise as a voltage and it is a series circuit. Given the loss and the maximum junction temperature, the required sink is a single calculation. Doing it is the difference between a working design and one that fails after ten minutes.

What you should now be able to explain or do

• Explain why a power device operates only fully on or fully off.
• Compare the four main device types on blocking voltage, current, speed and turn-off.
• Separate conduction loss from switching loss and say which one limits frequency.
• Explain why gate drive strength affects device temperature.
• Describe reverse recovery and say which device pays for it.
• Size a heat sink from the thermal path and the maximum junction temperature.

Check yourself