How to specify a custom solenoid: the 8 parameters you must define

Why this happens

Custom solenoids do not usually fail because the manufacturer got the magnetics wrong. They fail because the specification left something to be decided later by whoever built it, and that decision was made on the bench rather than against the application.

The pattern is consistent. Stroke and force get defined, because those are the two numbers everybody knows to ask for. The energisation profile gets compressed into a single duty percentage, because that is what the catalogue column shows. And what happens at the end of stroke gets left out entirely, because it is not obviously part of the magnetic question.

That last omission is the expensive one. A plunger of 8 g arriving at 0.5 m/s carries about 1 mJ of kinetic energy. One millijoule is nothing. Repeated 10⁶ times it is a kilojoule delivered into a pole face measuring a few tens of square millimetres, and because two nominally flat faces touch on a rim before they touch everywhere, it arrives on a fraction of that area. The face peens, the plunger can no longer close the gap fully, and holding force falls because force varies as the inverse square of gap. Recovering just 0.02 mm of lost closure on a 0.30 mm working gap costs about 12 % of the force, and it is a loss that accumulates rather than appearing. The relationship between gap and force is set out in how to calculate holding force.

The other omissions behave the same way: they move a decision from the specification to the bench. A duty figure of 100 % is only meaningful with an ambient temperature attached, because the limit is the permitted temperature rise, not the absolute temperature. A supply voltage without a tolerance range leaves the driver behaviour undefined. A mounting interface without a flatness figure leaves the working gap to be discovered during assembly, as covered in works on the bench, fails in the machine. None of these are exotic, and all of them are cheaper to state than to fix.

Check these in order

1. Stroke, with both ends defined. Not a single number. State where the plunger sits when de-energised, where it must be when energised, and what the useful force has to be delivered across. A device asked for 2 mm of travel between a 2.5 mm de-energised gap and a 0.3 mm energised gap is a different requirement from one asked for 2 mm with no stated end points.

2. Force at the worst point of the stroke, with its direction. Usually the start of pull-in or the end of stroke rather than the peak. Give the load as a curve if you have one: a load rising from 1 N at the start of pull-in to 4 N at the end is a different problem from a flat 4 N, even though both average differently against the same solenoid. Say whether the device pushes or pulls and how the load is coupled.

3. The real energisation profile. State it as four numbers rather than one: longest continuous energisation in seconds, operations per minute, total on-time per hour, and the ambient range the enclosure will actually see. Whether the plunger is seated throughout the long ones matters as much as the duty. This one parameter decides the wire gauge, and getting it wrong is the standard route to a coil burning out after a few weeks.

4. Supply and drive. DC or AC, nominal voltage and tolerance, available current at the worst case, and the drive method: direct switching, PWM holding, or a capacitor discharge. A ±10 % tolerance on 24 V puts the low limit at 21.6 V, and on DC that is about 81 % of the nominal force rather than the 90 % a linear reading would suggest, because force follows the square of current. If PWM is used, give the frequency and whether the coil sees the full supply at pull-in.

5. Environment. Ambient temperature range, condensation and washdown, dust, vibration, and any chemical exposure. A stated enclosure range of 0 to 40 °C is commonly 15 to 20 K higher at the winding, so the coil may see 60 °C, and the permitted rise is measured from there rather than from the room. The temperature range sets the hot-state force; the rest sets the sealing and surface treatment.

6. Mechanical interface. Mounting face flatness, bore concentricity, plunger coupling type, maximum side load, and the position and material of the end stop. Flatness of 0.05 mm across the pole area is a common figure and is worth writing down, because it appears directly in the working gap rather than as a separate tolerance.

7. Life target, and the definition of end of life. Cycle count alone is not sufficient. Say whether failure means a force loss of 10 %, a leak, a response time increase, or simply no movement. A life figure without a failure criterion cannot be tested. The mechanism also changes with the decade: 10⁵ cycles usually finishes without measurable wear, 10⁶ is where finish and friction start to matter, and 10⁷ needs the wear pair designed rather than inspected.

8. End-of-stroke impact, and who owns the deceleration. Allowed shock, whether a resilient bumper or a hard stop is used, and what the return spring is doing at that moment. If the specification is silent, the pole face is the brake.

What actually to change

Parameter left unspecified	What happens later	What to state instead
End-of-stroke impact	Pole face peens, gap opens, force falls over the life	Allowed impact energy, and the stop or bumper responsible
Duty cycle without ambient	Coil sized for 25 °C runs hot in a 60 °C enclosure	Duty plus ambient range plus longest single energisation
Voltage without tolerance	Force promised at nominal is not delivered at the low limit	Nominal, tolerance range, and worst-case available current
Working gap not defined	Assembly gap discovered on the bench, force short	Mounting flatness, concentricity, and the assembled gap target
Life without a failure criterion	Test cannot conclude pass or fail	Cycle count plus the measurable limit that defines failure
Return spring not specified	Spring fights the solenoid at the wrong end of stroke	Spring rate, preload, and force at both ends of travel
Side load left to the linkage	Friction consumes the margin, repeatability suffers	Maximum off-axis load and how it is guided
Coupling method not fixed	Plunger thread or pin distorts the armature	Coupling type, thread or pin size, and torque limit

When it IS the harder problem

The customer cannot give you a load curve. This is common, and the honest move is to measure rather than assume: instrument the linkage or the mechanism they are replacing and record force against position. A specification built on a guessed load curve is a specification that will be renegotiated.

The device has to fail safe. Then the de-energised state has to be the safe state, which means the return spring is a safety element rather than a convenience, and its force at end of stroke has to be guaranteed at the coldest temperature the machine will see. Cold stiffens grease and can stiffen some elastomers, so the worst case is usually not at room temperature.

The specification is a copy of an existing part. Reverse-engineering gives geometry and dimensions, not requirements. Establish the eight parameters anyway, because the reason the customer wants a new part is usually a failure the old part could not describe.

The unit sits in a system whose electrical behaviour you do not control. Shared supplies, long cable runs and controllers with their own ratings mean the coil may never see the nominal voltage. Then the specification has to state the worst-case supply rather than the nominal one, and the design has to hold its force there. A specification is only useful if it removes decisions rather than relocating them.