Relay circuits fail in the same surprising way over and over: the coil is fine, the switch is fine, everything reads correct on the bench — and the relay or the things around it keep dying. The usual culprit is the inductive kickback from the coil, and the usual fix is a flyback diode. But a diode in the wrong place is as good as no diode, and a diode with the wrong orientation actively makes the problem worse. This article covers why a relay coil produces up to a few hundred volts when it releases, where the flyback diode must go, whether a plain 1N4007 is enough, and the release-time trade-off that a faster circuit may force you to make.
Why a Relay Coil Needs a Diode: The 100 V Spike in One Sentence
A relay coil is an inductor, and an inductor resists change in current. When the switch opens, the current through the coil cannot stop instantly; it collapses over a short time, and the collapsing field induces a voltage that opposes the collapse — the back-EMF. In a typical 12 V coil with a reasonable switching speed, that induced voltage can spike to 100 V or more, far above the supply voltage and far above what the nearby circuit expects to see.
The spike is not a theoretical nicety. It stresses the switch contacts, arcing across them and shortening contact life; it couples into nearby sensitive electronics; and in an inductive circuit like a solenoid valve or contactor, it repeatedly applies 10× the supply voltage to components rated for the supply. The flyback diode is the standard cure because it gives the coil a closed loop to discharge into — instead of the voltage rising until something breaks, the current continues through the diode, and the spike never forms.
The “one sentence” version: the coil stores magnetic energy, and when the switch opens that energy has to go somewhere; without a diode, it goes into a voltage spike across the gap; with a diode across the coil, it goes back into the coil through the diode. Everything else in this article is about where that diode goes and how fast the release really is.
Placement and Polarity: Across the Coil, Cathode to Plus
The flyback diode sits across the coil, in parallel with it, with its cathode connected to the coil’s positive side and its anode to the negative side. In that orientation the diode is reverse-biased while the relay is energized — it carries no current and interferes with nothing. When the switch opens and the coil tries to push current in its old direction, the diode is forward-biased for that direction and shunts the collapsing current around the loop instead of letting it build a spike.
Two wrong placements are common and both defeat the purpose. Putting the diode in series with the coil instead of parallel blocks all current and the relay never works. Putting the diode back-to-front — anode on the coil’s positive side — makes it conduct permanently while the relay is energized, holding the relay contactlessly “on” through the diode and possibly frying it with continuous current. The polarity check is the same one used for any diode: the marking conventions and meter verification apply, and in a relay the rule is simple enough to remember — cathode to plus on the coil.
A second placement question is whether one diode per coil is enough. For a relay with a single coil, yes. For a contactor with multiple coils, each coil needs its own loop or a shared loop sized for the combined stored energy. And on the switching side, placing the diode across the coil rather than across the switch matters because the coil is where the energy lives; a diode across the switch does not stop the coil from spiking the rest of the circuit.
Rating the Diode: Coil Current and Supply Voltage Limits
The question every beginner asks is whether the 1N4007 in the parts drawer is enough, and the honest answer is usually yes — with conditions. The flyback diode must carry the relay’s hold current in the discharge loop and hold off the supply voltage in the forward direction. A 1N4007 handles an amp of continuous current and 700 V of reverse voltage with margin; small relay coils draw tens of milliamps to a few hundred milliamps, and supply rails run 5 V, 12 V, or 24 V. On that duty a 1N4007 is overkill but correct.
The conditions are where the small print lives. If the relay circuit is switched rapidly — opening and closing many times per second — the diode sees repeated discharge events, each of which stresses the junction thermally. The current per event is the coil current, but the events repeat, and a part that is fine for occasional switching may run hot in a high-repetition application. If the coil is a large solenoid drawing several amps, the diode should be sized to the coil current and the datasheet surge capability, not to the supply.
The temperature rule matters too. A diode rated 1 A at 25 °C carries less at 85 °C in a warm enclosure, so a high-duty solenoid next to a motor gets a smaller safe margin than the nameplate suggests. The general-purpose rectifier families cover the standard silicon choices for this duty, and the datasheet comparison method applies the same way it does for any power diode: read the rating at your operating temperature, not at the 25 °C headline.
Faster Release: Diode+Resistor, Zener, or TVS Configurations
The plain flyback diode makes the relay release slowly. Because the diode holds the coil current from collapsing instantly, the magnetic field decays over milliseconds, and the relay contact drops out with a distinct lag. For most purposes that lag is invisible. For solenoid valves, fast-switching timers, and time-critical automation, the lag is a real problem — and the fix trades release speed against spike suppression.
| Configuration | Spike suppression | Release speed | Complexity |
|---|---|---|---|
| Plain diode | Excellent (near-zero spike) | Slowest release | Lowest |
| Diode + series resistor | Good (smaller spike) | Faster | Low |
| Zener or reverse stack | Moderate (limited spike) | Faster | Medium |
| TVS | Moderate (clamps at rating) | Fast | Medium |
The diode-plus-resistor configuration is the classic middle ground: a resistor in series with the diode allows some current decay through the resistor, shortening release time while still capping the spike. The trade is that a larger resistor releases faster but lets the spike grow — the resistor has to be chosen for the application’s tolerance. The Zener and TVS configurations clamp at a designed voltage instead of near zero, converting some of the stored energy into a controlled spike that decays fast, at the cost of some heat and a finite energy rating per event.
The choice reduces to how fast the release must be and how much spike the circuit can tolerate. The repeated-event duty feeds the same logic the surge current ratings article applies to IFSM test waves, so a high-repetition relay should be checked against the datasheet surge curve rather than the steady rating. A pump valve that must slam shut in milliseconds gets a Zener or TVS; a generic relay in a light circuit is best served by the plain diode and its slow, gentle release. The protection family comparison covers the TVS-versus-Zener-versus-diode trade in more depth for the cases where the rail is sensitive and precision matters.
A worked pair makes the release-time trade quantitative. A 12 V relay with a 200 mH coil at 100 mA stores L·I²/2 = 0.5 × 0.2 × 0.01 = 1 mJ of magnetic energy. With a plain diode loop, that energy decays with the coil’s L/R time constant and the contact may take tens of milliseconds to release. Adding a series resistor that raises the effective loop resistance by a factor of ten shrinks the decay time by roughly the same factor, releasing the contact sooner — while the resistor-limited spike stays comfortably below the circuit’s insulation rating. The numbers are relief for the over-thinkers: a few ohms of resistance is often enough to halve the release time without any exotic parts.
AC Coils and Contactors: Why This Diode Does Not Apply
Everything above assumes a DC relay coil. An AC coil changes the rules completely, and the same diode placement that protects a DC relay will not work on an AC one. On AC the supply reverses every half-cycle, so a single diode across the coil would conduct during the half-cycle where it is forward-biased, holding the coil energized instead of protecting it, and burning out.
AC contactors instead use an RC snubber network or a different suppression method that works with the continuously reversing voltage. The concepts of limiting the back-EMF still apply, but the component is a resistor-capacitor network or a specialized device, not a flyback diode. If you are repairing or building an AC contactor, the DC-relay habit — one diode across the coil, cathode to plus — is actively wrong, and the correct path starts with the AC supply’s switching and the snubber sizing rather than the parts drawer’s diode collection.
The distinction also matters for universal-DC/AC relays that can be driven either way. Check the relay’s rating label before fitting a flyback diode: if the coil can be fed AC, the diode’s reverse-blocking behavior on one half-cycle invalidates it on the other, and the suppression approach must match the actual drive. The rule that survives every case is to read the coil’s supply type first, then apply the right suppression for that supply.