I built my own Wi-Fi connected irrigation controller using an Arduino Nano and some mechanical relays. The Arduino is opto-isolated from the business end of the relays which switch 24 V AC on/off to standard irrigation valve solenoids.

I seem to be "blowing up" the solenoid in the irrigation valve occasionally (too occasionally). When it fails, the solenoid in the irrigation valve is seen to be open-circuit (infinite ohms) versus the ~50 ohms I would usually see and the valve no longer operates.

I'm guessing that when I'm turning off the relay and cutting the current to the valve solenoid the inductive load is causing a big voltage spike and killing the solenoid. This is happening only on one set of valves which is located about 100 feet (30 m) from the controller which is probably adding additional inductive load through the long run of wires all bundled together.

Switching to a DC solenoid is a non-starter as 99% of all irrigation valves use AC. I'd rather not have to design my own irrigation valves. I also don't think a soft turn-on/turn-off is practical as this would require a new relay design.

I've been looking at a diode or snubber circuit to go across the solenoid to absorb voltage spike, but I am not too familiar with that. How can I proceed with this?

  • \$\begingroup\$ Farthest solenoid fails.... thinking indirect lightning might be a problem. Do you get it fail after a thunderstorm? \$\endgroup\$
    – Andy aka
    Commented May 8 at 20:13
  • \$\begingroup\$ My experience is that the driving circuitry will blow up way before the solenoid if the issue is lack of snubbing. \$\endgroup\$ Commented May 8 at 20:15
  • 1
    \$\begingroup\$ Please could you clarify your setup with a schematic? Do I infer correctly that you have DC relays (physically close to / integrated with the control system), which switch the AC supply to the solenoids (which are down potentially long cable runs)? \$\endgroup\$
    – Attie
    Commented May 8 at 20:42
  • \$\begingroup\$ I don't see how your control system would be different to any other (commercial) control system in the way that it operates these valve solenoids. If snubbers were necessary in commercial systems then they'd already be present in the valves. I think you're barking up the wrong tree here ... \$\endgroup\$
    – brhans
    Commented May 8 at 20:57
  • \$\begingroup\$ Difficult to make a schematic quickly. I am using an opto-isolated DC-controlled relay (3.3V) which is rated for up to 110V AC for the switch output. I agree with brhans. This does not seem that different from a commerical setup. I'm not sure whether a commercial system has diodes to protect against this. All I know is that in the past 2 months I've destroyed 3 solenoids...all the same way. Which is why I'm thinking its something to do with my controller and how it turns on/off the solenoid. There has been no lightening or power surges that I'm aware of in the past 2 months \$\endgroup\$
    – Jim M
    Commented May 8 at 23:03

6 Answers 6


Let me guess: the solenoids that keep burning out are the ones furthest from the controller out of all the others?

It's not voltage spikes, or lightning, or inductance on the lines, or anything like that.

Unlike DC solenoids where resistance limits the coil current, AC solenoids' current is limited by both the resistance and inductive reactance of the coil. However, the initial inductance is low because there is an air gap in the coil's center until the plunger/magnetic core is pulled down all the way into the coil.

For this reason, the inrush current for AC solenoids is typically several times (5-10 times) the holding current. But the length of this current spike is so brief that it doesn't have time to cause an appreciable amount of heating.

Except when your drive voltage is too low.

An AC solenoid is relying on that plunger getting into position quickly so the inrush current spike can be quick as well.

However, if your drive voltage is a little bit too low, then the plunger won't move into position fast enough and the inrush current spike is prolonged enough to blow the coil. Even if it causes a short, a short in the coil will very quickly give way to an open circuit failure from the wire acting like a fuse.

The rule of thumb is to not go more than 10% below the specified voltage for an AC solenoid for this reason. Your controller is probably losing a bit more voltage at the relay compared to the commercial controller, and when combined with the voltage drop from the 100 feet of wire between it and the solenoids, the voltage is probably just not quite high enough to reliably drive the solenoid. All it could take is slightly more force required to pull the plunger one time, or maybe it's a little too hot that day (increasing copper resistance) etc, and the plunger doesn't get into place in the coil fast enough and the inrush current kills the coil.

  • 2
    \$\begingroup\$ Good thinking. I had suspected inductance (in my answer) but hadn't considered the voltage drop which would be maximum when least desirable. \$\endgroup\$
    – Transistor
    Commented May 9 at 10:12
  • \$\begingroup\$ I have a question. Is this mitigated using a solid state relay with zero-crossing switch, such as the well known Omron G3MB-202P? I don't think so, since you are speaking about mechanical times, which I assume are longer than 20ms \$\endgroup\$
    – frarugi87
    Commented May 10 at 13:53
  • 1
    \$\begingroup\$ @frarugi87 I think use of a mechanical switch will improve your chances. The problem is the 0.2-2.0V drop in any solid state switch from the 24VAC budget when including your cable losses. A fix might be to use a 28VAC supply to compensate. \$\endgroup\$
    – KalleMP
    Commented May 11 at 10:58
  • \$\begingroup\$ This explanation sounds the most likely. A possible work around is to use thicker wire for the remote valves. \$\endgroup\$
    – KalleMP
    Commented May 11 at 11:04

If the solenoid were being killed by a voltage spike, it's far more likely there would be a short between turns, or to ground.

Applying 24 VAC for too long, dissipating ~12 W in the coil until it burns out might be the cause. Check that the coil is rated for 24 VAC continuous duty. Also check the the steady-state voltage across the coil is no more than 24 V.

Another possibility is electrolytic corrosion. This is not unlikely in a sprinkler system. If any water gets into the solenoid, the very fine wire in a coil will be "eaten" through.

  • \$\begingroup\$ In general these exact same solenoid valves have been operated with a commercial controller for years with no problems. So, the use of my controller is the only difference. I know these solenoids can run for 1-2 hours with no overheating (not hot to the touch), plus they are exposed to the air. The solenoid is in a sealed rubber/plastic casing which is designed for operation even in the rain, so I don't think this is a case of corrosion. \$\endgroup\$
    – Jim M
    Commented May 8 at 23:08
  • \$\begingroup\$ @JimM Are they exposed to sun? \$\endgroup\$
    – Antonio51
    Commented May 9 at 9:57
  • 2
    \$\begingroup\$ Is it possible the commercial controller drops the voltage after pull-in? I do that on most of the solenoid drivers I make at work... \$\endgroup\$ Commented May 9 at 14:09

Mechanical armature jamming and inductance

One thing that might be relevant:

When an AC solenoid is de-energised the armature is pushed out of the coil by a spring and pulled in by the energised coil.

In the deenergised position the solenoid's inductance (and its AC impedance or "resistance") is reduced. This has the beneficial effect that on turn-on a high current flows until the armature is pulled in. At that time the inductance increases and the current is reduced to a value sufficient to hold it (but not enough to pull it in the first place).

Perform a mechanical inspection on your failed solenoids. Look for torn or leaking diaphragms, water in the armature / coil, etc., and foreign material, mineral deposits that might gum up the movement.

If the armature can't move it is likely that the prolonged high current is overheating the coils. This can also be a problem with AC powered contactors used in industry. It would be very unusual for a coil to burn out due to a switching transient.

Related (but for a DC relay): https://electronics.stackexchange.com/a/487353/73158

  • \$\begingroup\$ I'll check that out, but as I said above, the solenoid is in a sealed rubber/plastic casing. I will have to cut them open to see the exposed wiring. The actuator on the end of the solenoid moves freely which makes me think it's not a mechanical problem. The inf impedance makes me think something burned out in the wiring. \$\endgroup\$
    – Jim M
    Commented May 8 at 23:11
  • \$\begingroup\$ Add a photo and part number to your question so we can consider potential failure modes. \$\endgroup\$
    – Transistor
    Commented May 9 at 10:09

In normal operation the valves will see the peak AC voltage on each half-cycle, this will be around 1.4 times the RMS so around 33-34V. When the relay opens the inductance of the coil and wiring will generate a large voltage spike, as you’ve identified. If you place two zener diodes in series with one reversed then this will limit the voltage. I’d start with about 40V devices, and with a current rating to match whatever the solenoid draws. Enhancements to this basic circuit could be a small series resistor to limit the current and a parallel capacitor to absorb high-frequency transients. Just using the two zeners should do the job well, the other components may be desirable if there are any EMC issues.

  • \$\begingroup\$ This is what I was thinking too. I was looking for a 36 to 40V Diode. I was thinking of putting them at the controller end. \$\endgroup\$
    – Jim M
    Commented May 8 at 23:01
  • \$\begingroup\$ It wouldn’t do any harm to put zeners at both ends \$\endgroup\$
    – Frog
    Commented May 9 at 6:21

when I'm turning off the relay and cutting the current to the valve solenoid the inductive load is causing a big voltage spike and killing the solenoid

The solenoid is a bunch of wire wound in a coil. It is about as hard to electrically destroy as that coil is. The primary ways to kill it are to:

  • Overheat it - it will then fail open,
  • Apply a transient overvoltage - the insulation on the coil wire will likely break down, and the coil will appear partially or fully shorted.

Switching to a DC solenoid is a non-starter

All solenoids are DC solenoids, more-or-less. They may need to run from a much lower voltage, though, since their AC reactance does not limit the DC current like it did for the AC current.

But why did you think you have to change to a DC solenoid?

Anyway, there are a couple of reasons why irrigation uses AC solenoid design:

  • the controller doesn't need a rectifier bridge nor a huge DC smoothing capacitor,

  • galvanic corrosion may be slower with AC than with DC

I've been looking at a diode or snubber circuit to go across the solenoid to absorb voltage spike.

You need to protect both the relay contacts from arcing, and the solenoid coil from overvoltage.

Thus, you need back-to-back Zener diodes or a TVS, either rated at 48VDC or so, both at the relay contact, and at the power entry to the solenoid.

  • 1
    \$\begingroup\$ I agree. I think its a voltage spike causing the insulation to break down. As per other suggestions, I will do an autopsy on one to see what I can glean from that. In the meantime, I need to buy some diodes. \$\endgroup\$
    – Jim M
    Commented May 8 at 23:13
  • 1
    \$\begingroup\$ @JimM That autopsy is key. \$\endgroup\$ Commented May 9 at 1:12
  • \$\begingroup\$ I threw together a SPICE simulation. From what I can tell on-line the inductance of the solenoid is in the range of 100mH and the impedance I've seen is around 50ohms. Modeling this as a resistor and inductor and adding a voltage controlled switch to turn on/off the 24VAC I am seeing a HUGE voltage spike on the order of kV's. I cannot say my simulation is very accurate, but it does say I would expect a large spike when turning off the 24VAC to the solenoid. Adding the zeners took care of this nicely. No spikes and the current looks reasonable. \$\endgroup\$
    – Jim M
    Commented May 9 at 4:16
  • \$\begingroup\$ I'm going to order some beefy Zeners (~47V or higher I think) and see how that works. Tomorrow I do an autopsy of a failed solenoid to see if its insulation breakdown or a popped solder joint. :-) \$\endgroup\$
    – Jim M
    Commented May 9 at 4:19
  • \$\begingroup\$ I have completed the autopsy of the failed solenoid. This was one well built device. I have to saw through a 1/8"+ of ABS casing, a 1/16" aluminum casing and another 1/16" of ABS to get to the coil. Took me about 2 hours of cutting. There is no signs of overheating (nothing melted or blackened), but I think I realize what the issue is. The lead wires to the solenoid are aluminum, and the coil is copper. Aluminum/copper connections are notoriously unreliable. The two metals are incompatible and create an oxide barrier. This barrier is susceptible to damage from electrical spikes. \$\endgroup\$
    – Jim M
    Commented May 10 at 21:45

It's (barely) possible that the spike is causing a short in the coil and your power supply is skookum enough to fry the coil to the point where it opens up. You can kill the spike with a bipolar TVS eg. P6KE56CA or 1.5KE68CA across the AC coil, and your relay will last longer. MOVs are possible too, but they do wear out and fail short with some degree of fireworks.

If this is the case you should be able to see that the coil has overheated. A similar effect would occur if your supply voltage was as bit on the high side, since the heating effect responds to the square of the voltage.

Or some kind of mechanical or electrolytic damage caused by design, manufacture, or installation.

Voltage spikes by themselves do not cause coils of wire to fail 'open'. If there is no evidence of overheating, I suggest looking elsewhere.

  • \$\begingroup\$ Note that MOVs degrade in proportion to energy absorbed; they wear in mains surge protection application where the energy levels are considerable, but you'd be hard pressed I think to even find a disc MOV that wears from such low power levels. \$\endgroup\$ Commented May 9 at 6:48
  • \$\begingroup\$ @TimWilliams I have seen multiple failures on mains-voltage relay coils, presumably similar energy. Relatively frequent switching though, several times per minute. \$\endgroup\$ Commented May 9 at 10:04
  • \$\begingroup\$ Thanks for the TVS recommendation! The bidirectional ones look perfect for this application, with plenty of guard band. \$\endgroup\$
    – Jim M
    Commented May 9 at 14:53
  • \$\begingroup\$ @SpehroPefhany Was the mains itself also surge limited prior? \$\endgroup\$ Commented May 9 at 16:53
  • 1
    \$\begingroup\$ @TimWilliams I believe the MOVs were across the contacts. \$\endgroup\$ Commented May 9 at 22:11

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.