Let's say I have a switched inductive load (e.g. a lift/hold electromagnet, electromagnetic actuator, or relay coil). IIUC, when the switch/EMR/SSR opens, the collapsing magnetic field can cause the switch to be exposed to some... nastiness.

AFAIUI, there are various "usual" ways to deal with this:

  • A lone flyback diode.
  • A flyback diode plus a Zener diode.
  • A TVS diode or similar "specialized" diode.

I think I sort of have a handle on the first two (with some caveats), but not so much on the third.

What are the various pros/cons of each approach, especially with respect to each of the specific inductor examples previously stated? And how does one select appropriate components for each of the above approaches? (Good answers should include guidelines such as "the flyback diode's reverse voltage must be at least that of the coil, and a margin of at least 2-3x is recommended" and should e.g. explain the difference in release time with and without a Zener and what applications are likely to be affected by that.)

p.s. Feel free to discuss other options (e.g. R+C, varistor), but as far as I can tell, their drawbacks make them sufficiently less desirable compared to diode-based solutions.


1 Answer 1



  • Minimizes the time it takes for the magnetic field to collapse.
  • Exhibits a near-infinite reverse voltage spike.

For relays, this option is the best for the relay as it minimizes the length of time during which contact arcing can occur. It may also exhibit the least contact bouncing. However, other components may not appreciate the reverse voltage spike. (In theory, given ideal wires, the spike is asymptotically infinite, but decays very quickly. In reality, wires aren't ideal, but the spike can still easily reach hundreds or even thousands of volts.)


  • Increases the time it takes for the relay contacts to open.
  • Must be installed in the proper direction.
  • Limits reverse voltage to near-zero.

This option is best for other circuitry, as there is essentially no reverse voltage. However, it is the worst for relay contacts as it results in slower opening and lengthens the time that arcing may occur, reducing contact life and potentially resulting in contact welding. Also, the slow collapse of the magnetic field may result in increased contact bounce due to the opposing magnetic and spring forces being close to balanced for a longer time.

For other types of inductors, however, this delay may (or may not) be a feature. For a lifting magnet or actuator, a "gentle release" may be preferred; however, it may also increase the likelihood of "sticking" if there is no force opposing the magnet. For a magnet opposing an irregular force (such as a magnetic door "lock"), a slow release may be fine.

TVS Diode

  • Less increase in contact open time compared to just a diode.
  • Limits reverse voltage to the clamping voltage.

This option strikes a good balance by allowing the reverse voltage to be limited to something that other components can withstand while minimizing the effect on coil open time.

The peak pulse ratings of many "typical" TVSs are hundreds of watts to kilowatts, with continuous power ratings in watts. A typical EMR that you might use on a PCB is likely to have a maximum power of maybe a watt, often less (e.g. 12V × 50mA = 0.6w). Thus, even cheap TVSs are massive overkill in terms of power rating. The voltage ratings are far more important. Standoff voltage needs to be higher than normal operating voltage, while breakdown voltage needs to be lower than what your components can handle.

Here, things get complicated, because the reverse voltage is a function of how much current passes through the TVS, and I don't know how exactly (other than testing) to determine how much reverse voltage will be seen for a particular EMR. You can easily "play it safe" by choosing a TVS whose maximum clamping voltage is below (perhaps with a 50% margin as well) what your other components can handle, but lower reverse voltage also means slower coil opening.

(Another) how do I choose an appropriate TVS for a relay? may have additional information. The key take-away, however, is that you are balancing protection for the relay (where higher reverse voltage is better) against protection of other components (for which lower reverse voltage is better). "Nothing" is best for the relay, while a flyback diode is best for other components. A TVS provides a better, balanced approach.

Diode + Zener Diode

  • Basically a TVS in two separate parts.
  • Must be installed in the proper direction, or can be duplicated to provide bi-directional protection.

Let's do some comparison, shall we?


simulate this circuit – Schematic created using CircuitLab

voltage plot current plot

Here we have voltage and current plots for a relay with a flyback diode, and a relay with a Z+D diode. The high Zener breakdown voltage (24 V) is meant to simulate a 24 V TVS diode. We can see that K2 "kicks back" a 24 V spike, which is limited by the Z+D pair (or TVS), but that the current through the coil drops very quickly. If the diode(s) are removed, the current, in the simulation, cuts off instantly and causes a voltage spike of ~20 TV! (You didn't have anything you cared about in that circuit, right?) By contrast, the current to K1 falls off much more slowly. In fact, if the cut-off current is 5 mA (or about 15% nominal voltage, which seems to be areasonable 'typical' value), K2 (Z+D) will open in ~1.7 ms and will be fully deenergized in just over 3 ms. By contrast, K1 (flyback) won't even open for over 8 ms.

Graphs are helpful, but this simulation provides an interactive demonstration of the difference between a flyback diode and a TVS/Z+D diode, as well as demonstrating the reverse voltage spike which occurs when a relay is deenergized. Start the simulation at high speed with the switch closed, wait for the relays to get close to their nominal current (~33 mA), then drop the simulation speed significantly and open the switch. Just as seen in the above graphs, the coil with the Z+D will jump instantly to ~24 V and will start to shed current quickly, matching the "cliff" that's visible in the above graphs, while the upper coil will lose current much more slowly and at a much lower ~600 mV.

As previously noted, this very small voltage is good for sensitive components downstream. However, it's worth noting that this also indicates that, while bigger is still better (from the perspective of draining the relay coil faster), almost any Zener or TVS, even with a relatively modest breakdown voltage, would be beneficial in terms of improving relay open time and reducing the chances of arcing. It is also worth noting that the current from a flyback diode alone follows a decaying-exponential falloff, which is the technical way of saying it drops more slowly over time. Since the opening current is likely to be relatively low compared to the nominal current, by the time the relay opens, the current will be dropping more slowly than initially. By comparison, the rate of current falloff with even a modest Zener / TVS is much closer to linear. (There is a slight slow-down as current approaches zero, but for the Z+D / TVS case, this doesn't have an appreciable effect until the current is already very small, on the order of microamps rather than milliamps.)


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.