Triac AC Load Switching Circuit Design

A triac switches an alternating load the way a relay does, but with no contacts to arc and no coil to drive. It also behaves quite differently from a relay: it latches on until the current falls to zero, it needs a gate pulse that is referenced to the line, and it has to be protected against the rate of change of voltage across it. Those three properties shape the whole circuit.

What a Triac Does

The device is two thyristors connected in inverse parallel on a single die, so it can conduct in both directions. Once triggered it remains conducting until the current through it falls below a threshold, which happens naturally as the alternating supply passes through zero, and then it blocks again until the next gate pulse.

That latching behaviour is what makes the triac simple to drive: a short pulse is enough, provided it is delivered at the right time and with enough current. It is also what makes the device unsuitable for direct current, because a direct current would keep it conducting indefinitely once triggered.

The voltage drop in conduction is about a volt, which is larger than the drop across a relay contact and produces heat. At ten amperes that is ten watts, which is enough to require a heat sink and a thermal design that a relay of the same rating would not need. The trade is the absence of moving parts and of contact wear.

Gate Drive and Triggering Quadrants

The gate has to be driven with a current relative to the terminal that is at the lower potential, and the polarity of that relationship changes as the supply reverses. The four combinations of gate and terminal polarity are called quadrants, and the sensitivity of the device differs between them, with the fourth quadrant usually the least sensitive.

gate current must exceed the specified trigger value for long enough for the device to latch, and the required value rises as the temperature falls. A design that just meets the typical trigger current will fail to fire at the cold end of the range, and the symptom is a load that works in a warm room and not in a cold one.

An optocoupler or a dedicated driver provides the isolation between the logic and the gate. The driver has to supply the trigger current with the correct polarity in both halves of the cycle, which usually means either two optocouplers or a device designed for the purpose. The gate circuit also has to be immune to false triggering from the rate of change of the load voltage.

Zero Crossing and Phase Control

zero crossing switching turns the load on only when the supply voltage passes through zero, which minimises the current surge and the electrical noise. It is the right choice for resistive loads such as heaters, where the only control needed is on or off, and for inductive loads where the inrush has to be limited.

Phase control switches the device on part way through each half cycle, which varies the power delivered. It is used for lamp dimming and for motor speed control, and it produces a waveform with fast edges that radiates interference in a broad frequency range. The filtering required for phase control is much more demanding than for zero crossing.

Choosing between them follows from the load. A resistive heater with a slow thermal time constant is better served by whole cycle control at a low repetition rate, which produces no interference at all, while a lamp that has to dim smoothly needs phase control and its associated filtering.

Holding Current and Load Compatibility

holding current is the minimum current that keeps the device conducting. If the load current falls below it during a half cycle, the triac turns off early, and the output waveform loses a portion of each half cycle. That produces a distorted waveform and, in a transformer coupled load, a large voltage spike.

Small loads are the usual problem. A relay coil or a small transformer may draw several amperes at the peak of the cycle and almost nothing near the zero crossing, so the device turns off before the end of the half cycle. A bleeder resistor across the load, which draws the holding current continuously, is the standard remedy at the cost of a small permanent loss.

Inductive loads also present a phase shift between voltage and current, and the device turns off when the current reaches zero rather than when the voltage does. The gate drive and the zero crossing detection both have to work with that relationship rather than against it, and the detection circuit is usually designed to fire at the voltage zero crossing, accepting the resulting asymmetry.

<img src="https://www.gopcba.com/wp-content/uploads/2026/08/pcb-clean-and-packaging.webp" alt="Triac gate drive and snubber network on a PCB” />

Snubber Networks and Voltage Rate

A triac can be triggered by the rate of change of the voltage across it even with no gate signal, because the internal capacitances couple a transient into the gate structure. A fast rising voltage, such as the one produced when an inductive load is switched off, can therefore turn the device on when it should be off.

A snubber network, usually a resistor and a capacitor in series across the device, limits that rate of change. The capacitor absorbs the fast transient and the resistor damps the resulting oscillation, and the values are chosen from the device specification and from the load. A snubber that is too small does nothing, and one that is too large passes a continuous current that heats the resistor.

The placement of the snubber matters as much as its value. It must be connected directly across the terminals of the device with the shortest possible leads, because the inductance of a longer connection prevents the capacitor from responding to a fast transient. A snubber placed at the load end of the wiring does not protect the device at all.

Devices described as snubberless tolerate a higher rate of change without external components, and they can also handle loads with a poor power factor. They cost more than a standard triac, and the choice follows from the load rather than from the price.

Isolation and Safety

The gate circuit is referenced to the live side of the supply, so the logic that controls the relay has to be isolated from it. An optocoupler or an isolated driver provides that barrier, and the layout of the barrier follows the same rules as any other isolation: a defined creepage, no copper crossing the gap and no test point that bridges it.

The creepage and clearance requirements follow from the working voltage, which is the mains voltage rather than the logic supply. On a board that also carries low voltage circuitry, the mains area should be a physically separate region with the isolation barrier marked on the assembly drawing, and the mechanical design should prevent a user from touching the mains side.

Track spacing in the mains area has to satisfy the applicable standard for the pollution degree of the environment. A coated board may allow smaller spacing than an uncoated one, but the credit only applies if the coating is applied correctly over the whole area, including under components. Assuming the credit without the process is a common cause of a failed safety review.

Layout and Thermal Design

The load current flows through the device and through the traces on both sides, so those traces have to be sized for the current and for the temperature rise the insulation can tolerate. A triac dissipating ten watts in a small package needs a copper area large enough to move that heat to the air, and the general arrangements for that are described in PCB thermal management.

The gate circuit should be kept short and away from the load traces. A gate trace that runs parallel to the load current path picks up a voltage proportional to the rate of change of the load current, and that voltage can trigger the device. The general switching layout rules in switching regulator layout apply to the loop formed by the device, the snubber and the load.

Clearance around the device is part of the thermal design as well as the safety design. A metal heat sink connected to the tab is live unless the device is isolated, and a heat sink that is earthed while the tab is live requires an insulating pad. Both the safety and the thermal consequences of that choice have to be considered together, and the cost of the different options is discussed in manufacturing cost reduction.

Verification and Common Faults

Verify the design with the intended load rather than with a resistive test load, because the behaviour of an inductive load is quite different. Measure the voltage across the device at turn off and confirm that the peak stays within the rating, and measure the current waveform to confirm that the device stays conducting for the whole half cycle.

A device that misfires when a nearby contactor operates is being triggered by a voltage transient rather than by its gate. Increasing the snubber capacitor, or moving the snubber closer to the device terminals, is usually the fix, and comparing the behaviour with the snubber fitted and removed confirms it.

A device that runs hot at a current well within its rating usually has a gate drive that is marginally low, which keeps the device in a partially conducting state for part of each cycle and increases the dissipation. Checking the gate current at the cold temperature extreme is the way to find it, and the practice of verifying parameters at the extremes rather than at room temperature is the same one used throughout production testing.

FAQ

Can a triac switch a DC load? No. Once triggered, a triac remains conducting until the current falls to zero, which never happens with a direct current. A transistor or a MOSFET is used instead.

Why does my triac turn off in the middle of a half cycle? The load current has fallen below the holding current. Add a bleeder resistor across the load to keep the current above that value.

Do I need a snubber on a resistive load? Often not, but the wiring inductance can still produce a fast transient. A small snubber is cheap insurance on any mains switching circuit.