Switching a solenoid, contactor coil, or relay from a digital output looks harmless, but every time that output opens, the coil fights back with a high-voltage spike that pits contacts, stresses solid-state switches, and throws electrical noise across the panel. Do it thousands of times and you get pitted relay contacts, mysteriously failing output cards, and intermittent glitches nobody can pin down. A snubber - or for DC, a simple flyback diode - tames that spike. This guide explains where the spike comes from, which suppression to use for DC versus AC loads, where to mount it, and how the missing-snubber symptoms show up in the field.
Output Snubber in one line: An output snubber is a small suppression component placed across an inductive load - a solenoid, contactor, or relay coil - to absorb the high-voltage spike that the coil produces when the output switches off. For DC loads the usual device is a flyback (freewheeling) diode; for AC loads it is an RC snubber or a metal-oxide varistor (MOV). Without it, the inductive kickback arcs across relay contacts and stresses solid-state outputs, causing pitted contacts, electrical noise, and premature output-card failure. The suppression is mounted at the coil, not at the output card.
An inductive load - anything with a coil, such as a solenoid valve, a motor contactor, or a control relay - stores energy in a magnetic field while current flows through it. That stored energy is the root of the problem. A coil resists changes in the current through it, and when the output opens the circuit, the current cannot stop instantly; the collapsing magnetic field forces the coil to generate whatever voltage it takes to keep the current flowing for an instant. Because the circuit has just been broken, that voltage climbs extremely high - far above the supply voltage, into the hundreds of volts even on a modest 24 V coil - as the coil dumps its stored energy into the newly opened gap. This is the inductive kickback, also called back-EMF or the flyback spike.
That high-voltage spike has to go somewhere, and where it goes is the output that just opened. On a relay output, the spike jumps the gap between the separating contacts as an arc, and that arc is what erodes the contact metal a little more with every operation. On a solid-state output, the spike appears across the semiconductor switch, and if it exceeds the device's voltage rating it stresses or destroys it. The spike also radiates electrical noise into nearby wiring, which can corrupt analog signals, upset communications, or cause phantom input transitions. All of this happens in a fraction of a millisecond at the instant of turn-off, invisibly, every single time the coil is switched - which is why the damage is cumulative and the eventual failure looks so mysterious.
For a DC-powered coil the standard and highly effective suppressor is a flyback diode, also called a freewheeling or catch diode, wired across the coil reverse-biased so it does nothing while the coil is energized. The moment the output opens, the collapsing field reverses the coil's voltage, forward-biasing the diode and giving the current a harmless loop to circulate through as it decays, instead of forcing an arc across the output. This clamps the spike to roughly one diode drop above the supply and protects the contacts or the semiconductor almost completely. The one tradeoff is that a plain diode slows the coil's release slightly because it lets the current decay gently; where fast dropout matters, a diode combined with a zener or a resistor speeds the decay while still clamping the spike.
AC loads cannot use a plain diode because the coil voltage alternates, so a diode would conduct on every half cycle. The AC solutions are the RC snubber and the metal-oxide varistor. An RC snubber - a resistor and capacitor in series, placed across the coil or the contacts - absorbs the switching transient into the capacitor and damps it with the resistor, softening the spike and quieting the arc. A metal-oxide varistor is a voltage-clamping device that stays effectively invisible below its threshold and then conducts hard to clamp any voltage that tries to exceed it, catching the spike's peak. The two are often used together, the RC snubber to damp the transient and the varistor to clamp the peak. Some output cards come with a degree of built-in suppression, but for a significant inductive load, external suppression sized to the coil is what actually protects the output over a long service life.
Where you put the suppression matters as much as choosing it. The correct place is right at the inductive load - across the coil itself - not back at the output card. Suppressing at the coil kills the spike at its source, so the high voltage never travels back along the field wiring toward the card and never radiates noise from that whole length of cable. If you mount the suppressor at the card instead, the spike still races down the wire between the card and the coil, arcing and radiating along the way, and you have protected the least of the problem. Mounting at the coil is a small discipline that pays off in both contact life and a quieter, cleaner panel. The rule is simple: suppress the source, and the source is the coil.
The absence of proper suppression leaves a recognizable trail that remote monitoring helps read. Pitted, high-resistance relay contacts that slowly stop making reliably; output cards that fail earlier than they should and get blamed on bad luck; and intermittent glitches on nearby analog or digital channels that flicker exactly when a particular coil switches - these are the fingerprints of missing snubbers. A cloud SCADA platform such as Merobix historizes both the output commands and the surrounding channels with timestamps, so an engineer can line up a mysterious analog blip or a phantom input transition against the moment a specific solenoid de-energized and see the cause staring back. Catching that correlation from the office turns a baffling, drawn-out troubleshooting effort into a clear conclusion - add suppression at that coil - and heads off the slow contact erosion and card failures that unsuppressed inductive switching would otherwise keep producing.
At the coil. Mounting the suppression right across the inductive load kills the voltage spike at its source, so it never travels back along the field wiring toward the card and never radiates noise from that length of cable. If you mount it at the card instead, the spike still races down the wire between the card and the coil, arcing and radiating on the way, so you protect very little. The rule is to suppress the source, and the source is the coil.
A flyback diode is used on DC loads: wired reverse-biased across the coil, it gives the collapsing coil current a harmless loop to decay through, clamping the spike to about one diode drop above the supply. An RC snubber is used on AC loads, where a diode cannot be used because the voltage alternates; the resistor-capacitor pair absorbs and damps the switching transient. AC loads often also use a metal-oxide varistor to clamp the peak, sometimes alongside the RC snubber.
Every time the output opens, the coil produces a high-voltage kickback that arcs across relay contacts or stresses a solid-state switch, and it radiates electrical noise into nearby wiring. Over thousands of operations this erodes and pits relay contacts until they stop making reliably, shortens the output card's life, and can cause intermittent glitches on neighboring analog and digital channels. The damage is cumulative and invisible in the moment, which is why the eventual failure often looks mysterious until someone connects it to the unsuppressed coil.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.