RS-422 Interface Circuit Design

RS-422 is a balanced interface that carries a single signal in one direction over a pair of conductors. It is the simplex counterpart of the bidirectional bus that most people know, and it is still used where a one way link has to travel a long distance through a noisy environment.

The Electrical Basis

A driver produces complementary signals on the two conductors of the pair, and the receiver responds to the difference. Noise that couples equally into both conductors is subtracted at the receiver, which is where the noise immunity comes from.

The standard specifies a differential voltage of at least two volts into a terminated load and a common mode range over which the receiver must work. Those two numbers define the interface, and every part that claims compliance meets them.

The pair must be balanced for the rejection to work. Equal track lengths, equal impedances and a shared return path are the conditions, and a layout that treats one conductor as a signal and the other as a return loses most of the benefit.

Drivers and Receivers

A driver has a differential output that is either a pair of complementary transistors or a bridge. It must survive a short between the two outputs, and the common parts include a current limit for that reason.

A receiver has a differential input with a high input impedance, and its threshold is centred near the middle of the common mode range. A receiver with a wide common mode range is chosen when the two ends are on different earth systems.

One driver can feed several receivers, and the standard specifies how many unit loads may be connected. The number is limited by the leakage and the capacitance of the receivers, which is the same reasoning as on any shared bus.

termination and Reflections

termination matches the impedance of the cable so that the signal is absorbed rather than reflected. A twisted pair of about a hundred and twenty ohms is terminated with a resistor of that value at the far end of the line.

A second termination at the driver end is not required for a one way link, and adding it doubles the current the driver supplies. Where the driver is connected to a long stub, a series resistor at the source is used instead to damp the reflection.

The termination has to be at the physical end of the cable. Placed at a connector with another metre of cable beyond it, the resistor does nothing useful and the reflection continues.

cable length and Data Rate

cable length and data rate trade against each other, because the loss of the cable rises with frequency and the received amplitude falls. A ten megabit link reaches a few metres, while a hundred kilobaud link reaches more than a kilometre.

The limitation is the attenuation at the highest frequency component of the signal, not the clock rate itself. A signal with fast edges has more high frequency content and suffers more, so a driver with a controlled slew rate extends the achievable distance at the cost of a lower maximum rate.

Skew between the two conductors of the pair also matters at high rates. A difference in length of a few centimetres turns part of the differential signal into a common mode signal, which is then rejected along with the noise.

<img src="https://www.gopcba.com/wp-content/uploads/2024/10/pcb-coloro.jpg" alt="Differential pair termination and driver layout on a PCB” />

common mode range in Practice

The common mode range is the window of average voltage on the pair within which the receiver still works. It is quoted with respect to the local ground of the receiver, and it has to cover the ground potential difference between the two ends.

That difference is what causes most field problems. Two buildings on different earth systems can differ by several volts, and a momentary difference during a fault can be much larger.

An isolated receiver, or a pair of optocouplers on the logic side, removes the limit entirely. The isolation is then the parameter to check, and its common mode transient immunity has to exceed the rate of change of the difference during a fault.

Protection

The interface leaves the enclosure, so it sees electrostatic discharge and surge. A transient suppressor at the connector, followed by a small series resistance, limits what reaches the driver and the receiver.

The series resistance also decouples the capacitance of the suppressor from the line. Without it, the capacitance loads the pair and reduces the achievable data rate.

Where the cable runs outdoors, a gas discharge tube or a metal oxide varistor handles the bulk of a lightning induced surge, and the suppressor diode catches the leading edge. The arrangement follows the usual layered pattern.

Layout and Grounding

Route the two conductors as a pair with equal length and constant spacing, and keep the reference plane continuous underneath them. A break in the plane changes the impedance and produces a reflection.

Keep the differential pair away from switching circuits and from other pairs. Coupling from a neighbouring pair appears as a differential signal and cannot be rejected, which is the failure mode that surprises designers most often.

The logic side of the driver shares a ground with the controller, and the cable side is referenced to the connector and the chassis. The two meet at one point, and the general approach is the same as for any mixed domain board, described in our guide to mixed signal board design.

Multi Drop and Repeaters

One driver can feed several receivers on the same pair, and the limit is set by the total leakage and capacitance of the nodes rather than by the standard alone. Beyond that limit a repeater is used, which receives the signal and drives a fresh pair.

A repeater adds its own delay and its own failure mode. On a long link the delay belongs in the timing budget of the protocol, and a repeater that fails leaves the far section of the network without data and without an obvious symptom at the near end.

Verification and Faults

Verify the link with an oscilloscope at both ends and with the receiver terminated. The differential amplitude at the far end is the number to record, and comparing it with the driver output shows the loss of the cable.

Measure the common mode voltage at the receiver with the system running. A value near the edge of the specified range indicates a grounding problem rather than a cable problem.

An intermittent link usually has a termination that is missing or misplaced, or a pair that is split at a connector so that the two conductors take different routes. The release checks that keep such an interface consistent are collected in our PCB design release checklist, the assembly points in judging PCB quality, and the measures that control the edges of a fast signal in our guide to high speed PCB design rules.

FAQ

Do I need a termination at both ends of an RS-422 link? Only at the receiving end for a one way link. A second termination doubles the driver current and adds little.

How far can RS-422 reach? More than a kilometre at a low data rate, and a few metres at ten megabits. The limit is the attenuation of the cable at the highest frequency in the signal.

Why does the link fail between two buildings? Ground potential difference. Isolate the receiver or use a differential link with a wide common mode range.