Solenoid coil burnout after weeks: the real duty cycle
A coil that survives a month and then dies is being thermally overloaded, and the percentage duty cycle in the specification is usually not the one running. Current-clamp a full machine cycle and time the on-period yourself — a startup branch, a latched fault output, or an operator override can turn a 50% duty into 70% without anyone changing a drawing.
Why this happens
Coil failures divide cleanly by how long they took, and the timing tells you which family you are in. A coil that fails in minutes or hours has an electrical problem: wrong voltage, a damaged turn, a short to the housing, a supply that is not what the label says. A coil that fails after weeks or months has a thermal problem. It is not being destroyed, it is being cooked, and the average heat input is exceeding the average heat removal.
Percentage duty cycle is the number that governs this, and it is the number most often wrong. It is defined as on-time divided by total cycle time, and it is only meaningful together with the ambient temperature the coil sits in. A coil rated 50% duty at 40 °C in a ventilated housing does not have a 50% duty allowance at 55 °C inside an enclosure. The allowance shrinks as the ambient rises, and it shrinks again if the supply sits high.
The reason this catches so many machines is that the nominal cycle is genuinely 50%, and the exception is somewhere in the control logic. Nobody lied. A startup hold, a latched fault output, a manual override, or a commissioning mode can each add a long energised period that appears once per shift and never in the specification.
Check these in order
1. Current-clamp the coil and record several complete cycles. Include power-up, include at least one shift change, include whatever the operator does at the end of a run. Do not sample for thirty seconds and extrapolate. You are looking for the longest single on-period, because that is what sets the peak temperature.
2. Compute duty from what you recorded. On-time divided by cycle time. Then do it again with the longest observed on-period instead of the nominal one. The gap between those two numbers is usually the whole story.
3. Measure the ambient inside the enclosure, running, on a warm day. Not the workshop. The interior of a working machine can sit 20 to 30 K above the room, and that difference comes straight out of the allowed rise.
4. Check the supply voltage at the coil under load. Copper loss goes as voltage squared over resistance. Ten percent high is about 21% more heat. Twenty percent high is 44%. A winding designed to a tight margin will not survive that, and it will take months rather than minutes to show.
5. Check whether the hold is PWM or continuous. If the machine holds the coil with PWM, the relevant quantity is the RMS current during the hold, not the peak. If it holds continuously, the relevant quantity is the on-time and the thermal time constant. Mixing the two up is common: a PWM hold gets analysed as if it were continuous, or a continuous hold gets defended by pointing at a low average duty.
6. Check the thermal path before concluding the coil is undersized. A potted coil in a closed plastic housing with no conduction path to the frame will run hot even at a modest duty, because there is nowhere for the heat to go. Adding copper is expensive; giving the heat a shorter path is often free.
What actually to change
| Finding | What to change | Why not the other thing |
|---|---|---|
| One long hold per cycle | PWM hold, or a mechanical hold that draws nothing | Removes the heat source instead of fighting it |
| Ambient inside enclosure high | Ventilate, relocate, or thermally isolate | A higher insulation class only buys time |
| Supply running high | Regulate the supply, or specify the coil for the real voltage | Over-voltage heats a coil quadratically |
| Duty genuinely above rating | Enlarge the copper window or step to a larger frame | More turns of finer wire makes it worse |
| PWM hold analysed as continuous | Compute RMS current during hold | Peak current is the wrong basis in PWM |
| Thermal path blocked | Provide a metal path from coil to frame | Insulation upgrade does not fix a bottleneck |
When it IS the harder problem
A hold that is short but frequent. Percentage duty treats all on-time as equal, but the winding has thermal mass, so a 5-second pulse every 10 seconds is not the same insult as a 300-second pulse every 600 seconds, even though both are 50%. The difference is whether the winding reaches equilibrium during each pulse. For pulses much shorter than the thermal time constant you can run a higher percentage than the rating suggests; for pulses much longer, the rating is the rating. This is why the same coil survives one machine and dies on another with an identical duty figure on paper.
Heat that never leaves the winding. The copper loss is the only heat source, but copper is rarely the bottleneck. The bottleneck is whatever sits between the winding and the outside world. A potted coil in a closed plastic housing with no metal contact to the frame accumulates heat in the winding while the outside of the housing still feels merely warm. Surface measurements then reassure everyone that the coil is fine, right up until the insulation gives up.
A failure at month three that was a decision at design time. Most of the coils I have seen die on time did so because a margin was spent somewhere early: an ambient assumed at 25 °C, a supply assumed nominal, a duty assumed nominal, a housing assumed to conduct. Each assumption is reasonable alone. Multiplied together they give a mean winding temperature 30 K above what anyone calculated, and insulation degradation runs on a steep curve with temperature. The failure is real, but its cause is a stack of papers rather than a part.
A note on what this page is
This is a personal notebook, not a product page. I write down the checks that actually decide the outcome and the order to run them in, including the ones I got wrong first.
solenoid coil burnoutsolenoid duty cyclesolenoid coil overheatingsolenoid percentage dutysolenoid coil failure
Frequently asked
- The coil lasted a month, so it was not a duty cycle problem, right?
- The opposite. A month is the signature of a thermal overload rather than an electrical one. A wrong voltage or a shorted turn kills a coil in minutes or hours. A coil that dies after weeks is being cooked slowly, which means the average heating exceeds the average cooling over a long period.
- The specification says 50% duty and the machine runs 50%. What else is there?
- The specification is about the nominal cycle. Check the exceptions: a startup sequence that holds the valve longer, a fault branch that latches the output, a manual override an operator leaves engaged, a test mode nobody documented. Current-clamp the coil for several complete cycles including startup and include the worst case in your duty calculation.
- Does a short pulse at high current count against duty cycle?
- Less than you would expect, because the winding has thermal mass and a short pulse does not have time to heat it fully. Percentage duty is the steady-state approximation; for pulses much shorter than the winding thermal time constant you can exceed the nominal figure without reaching the same temperature. The trap is treating a long hold as if it were a series of short pulses.
- The supply voltage is 10% high. Does that matter for heating?
- More than proportionally. Copper loss goes with voltage squared over resistance, so a 10% voltage excess is about 21% more heat in the winding. If the coil was designed with minimal margin, that is enough on its own to explain a failure at month three rather than year three.