ESD Protection Layout Design Guide

An electrostatic discharge is a current pulse of tens of amps with edges measured in hundreds of picoseconds. No device survives it by absorbing the energy; the circuit survives by giving the current a short path to ground that bypasses everything sensitive.

What the Discharge Looks Like

The human body model produces a pulse with a rise time of a few nanoseconds and a duration of a hundred. The current is limited by the body resistance, and the contact discharge in the standard model is faster still.

The peak current of a few amps flows for a few nanoseconds, and the voltage across an inductance of just a few nanohenries during that time is tens of volts. That is why the length of the ground return is the parameter that decides whether the protection works.

The event can occur at any point a person can touch: a connector, a switch, a display bezel or a metal part of the enclosure. Each of those paths has to be considered separately.

Clamp Diodes and Where They Go

A clamp diode conducts when the voltage exceeds its threshold and diverts the current to a rail or to ground. It has to be closer to the source of the discharge than the circuit it protects, which means at the connector.

A clamp placed a centimetre down the board leaves the trace ahead of it to carry the current, and the voltage developed across that trace appears at the protected device. The protection is then theoretical.

Two diodes, one to the positive rail and one to ground, are the classic arrangement for a signal line. A single bidirectional device replaces them where the signal swings both ways, and an array protects several lines with a common ground connection.

The ground return Path

The current has to reach the reference of the discharge, which is usually the chassis or the earth of the equipment under test. On a board that means the connector shell or a dedicated protective ground, not the signal ground.

A long thin trace from the clamp to the signal ground has inductance, and the voltage developed across it appears between the protected device and its own reference. The device then sees a difference of tens of volts even though the board is protected in theory.

The protective ground should be a plane or a wide area, and the clamps should connect to it with the shortest possible traces. Where the chassis and the signal ground are joined, the joint should be at one point near the connector.

Connector and Creepage Design

The connector shell should be earthed to the chassis, and the signal pins should be placed so that a discharge from a finger reaches the shell first. That is a mechanical design decision as much as an electrical one.

The creepage between the pins of a connector is small, and a discharge can jump between them. Choosing a connector with a larger pitch, or with a grounded pin between signal pins, reduces the risk.

Where a connector has a metal shell, bonding it to the chassis with several points reduces the inductance of the discharge path. A single wire is much worse than a direct metal to metal contact.

Guard Rings and Board Features

A guard ring around a sensitive node intercepts the surface current of a discharge. The ring must be connected to a low impedance reference, and a ring that is floating makes the situation worse rather than better.

Slots in the board under a connector increase the surface path and can break a creepage path. They are used where the standard requires a distance that the layout cannot otherwise achieve.

A spark gap is a deliberate pair of exposed copper features with a defined separation, placed so that a discharge jumps there rather than into the circuit. It works and its trigger voltage varies with humidity and contamination.

Protection of the Supply Rails

A discharge that reaches the supply rail through a clamp diode raises the rail by tens of volts. The decoupling capacitors absorb part of it, and a transient suppressor across the rail clamps the rest.

The rail clamp has to be faster than the regulator, which cannot respond in nanoseconds. A small ceramic capacitor close to the clamp and a suppressor with a low dynamic resistance are the combination used.

Where the rail supplies a sensitive analog circuit, a series ferrite and a further capacitor isolate that circuit from the transient. The decoupling is then a filter rather than a single capacitor.

Selecting the Protection Component

The working voltage and the capacitance of the clamp are the first two parameters. A signal that swings to five volts needs a device with a standoff above that value, and a high speed line needs a capacitance of a picofarad or less.

The clamping voltage at the peak current is the third. A device that clamps at forty volts protects a fifteen volt circuit and does nothing useful for a five volt one, so the clamp is chosen for the device that follows rather than for the connector.

The dynamic resistance of the clamp comes from the published curve, and the voltage at the protected node is the clamp voltage plus the product of the current and the impedance of the path. Considering only the headline clamping figure is a common and expensive error.

Product Level Considerations

The enclosure is where the discharge enters, and the route it takes inside the product depends on the internal layout. A plastic enclosure with a metal chassis inside behaves quite differently from one that is entirely plastic.

A grounded internal metal part gives the discharge a preferred path that can be arranged to avoid the board. Where no metal part exists, the board itself has to carry the current, and the connectors become the entry points that the layout has to protect.

Testing and Verification

Verify the design with a discharge generator at the levels required by the relevant standard, applied to the points a user can touch. The test is repeated with both polarities and at several points on the enclosure.

A failure appears as a reset, a lock up, a corruption of memory or a permanent fault, and each has a different cause. A reset that occurs at a high test level indicates a transient on the supply, while a permanent failure indicates a device that was not protected.

Record the level at which each symptom appears, and change one thing at a time. The transfer of the discharge path is often the effective change, and adding a component without improving the path rarely helps. The release checks that keep such protection in place are collected in our PCB design release checklist, the assembly points in judging PCB quality, and the wider measures in our guide to EMI suppression design principles.

FAQ

Why did my protection diode not save the circuit? It was too far from the connector. The trace between the connector and the diode carries the discharge into the board.

Should the connector shell be earthed? Yes, to the chassis with a short and wide connection so that a finger discharge reaches it before the signal pins.

Do I need a spark gap as well as diodes? Usually not. A spark gap is a fallback where the standard demands a path that the layout cannot provide otherwise.