How to calculate solenoid holding force (and the correction everyone skips)
Holding-force equations assume ampere-turns stay constant, and they do not. Copper resistance rises 0.393 %/K, so a winding stabilising 70 K above ambient sits at about 78 % of its cold current and about 61 % of its cold force. Here is the order to run the calculation, with the gap, fringing and saturation corrections that actually move the answer.
Why this happens
Every closed-form route to holding force comes back to the same two expressions. From flux density, F = B²A / 2μ₀, where A is the pole face area and μ₀ is 4π × 10⁻⁷ H/m, or 1.257 × 10⁻⁶ H/m. From the winding, F = (NI)² μ₀ A / 2g², where NI is the ampere-turns and g is the working air gap.
Both are correct within their assumptions, and they share one built in: that NI is a number you already know. It is not. It is the current the winding actually draws at the temperature it actually reaches, and at constant voltage current is set by resistance.
The temperature correction is the one almost nobody applies. Copper resistance rises at 0.393 % per kelvin. A winding measured at 20 °C that settles at 90 °C has gained 70 K, so its resistance is 27.5 % higher and its current is down to about 78 % of the bench value. Force goes with the square of current, so it arrives at roughly 61 % of the figure you calculated. That is not a rounding error, and it is why a design with only a thin margin at ambient is genuinely marginal inside a warm enclosure. The mechanism is set out in more detail in why force drops when the coil gets hot.
The gap in the formula is not the gap on the drawing. A pole face specified 0.30 mm from its stop may sit 0.35 mm away once mounting face flatness, plunger perpendicularity and stop wear are stacked in. The relationship is inverse square while unsaturated, so that 0.05 mm is a 16.7 % increase in gap and a 26 % loss of force. This is the single largest error source in the whole calculation, and it is also the one that is invisible until you measure a real assembly. Mounting geometry problems are covered in works on the bench, fails in the machine.
Two corrections then pull in opposite directions. Fringing extends the effective pole area by roughly one gap length around the perimeter. On a 6 mm face the area is 28.3 mm² and the perimeter is about 18.9 mm, so a 0.30 mm gap adds around 5.7 mm², which is a 20 % increase in effective area and therefore in force. Leakage subtracts: a leakage coefficient of 1.5 to 2.5 is normal for an open frame, against 1.1 to 1.3 for a tubular design, and because force varies with the square of flux density a leakage coefficient of 2 quarters it. Small closed designs get the fringe benefit and little of the leakage penalty; open frames get the reverse.
And the ceiling is saturation. Low-carbon steels run out of flux somewhere around 1.6 to 2.0 T. Past that point the pole face is saturated, extra ampere-turns produce heat rather than flux, and the force curve flattens while the current curve keeps climbing.
Check these in order
1. Fix the working gap before anything else. Start from the assembled gap, not the nominal one. Add mounting face flatness, plunger perpendicularity, plating thickness, stop wear allowance and any deliberate non-magnetic shim. Then sanity-check the number against a physical measurement on one unit.
2. Get the cold ampere-turns from the supply and the measured resistance. Use measured resistance, not the nominal figure, and include the lead wires. A 5 % error here becomes a 10 % error in force.
3. Apply the temperature correction to current, then square it. Estimate the stabilised winding temperature from the duty cycle and the enclosure, not from the ambient. The 0.393 %/K coefficient applies over the whole span.
4. Compute force with a fringing-corrected area and a defensible leakage coefficient. Take the coefficient from the topology, not from optimism. Tubular encloses its own return path; an open frame radiates and leaks.
5. Check the working point against saturation. If doubling the current does not roughly double the force in your model, you are already past the knee and the model has stopped being useful. Saturation is diagnosed from data in finding saturation from measurements.
6. Subtract the return spring at the end of stroke, not at the start. A compression spring is at its highest force exactly where the air gap is smallest and the solenoid has least left to give. Compare hot holding force against worst-case spring force, and require real margin between them.
7. Calibrate against one measurement and scale the rest. Measure at one voltage, one gap and one temperature, compare with the model, and carry the ratio through the remaining conditions. This closes more error than any amount of extra arithmetic.
What actually to change
| Symptom in the calculation | What to change | Why not the obvious thing |
|---|---|---|
| Force marginal once hot | Reduce the working gap, or increase wire diameter and accept the extra current | Changing core material typically moves force by only a few percent |
| Gap larger when assembled than on the drawing | Tighten flatness and concentricity, or add a positive stop | Grinding the plunger moves the stop rather than the gap |
| Force flat while current keeps rising | The pole face is saturated; reduce ampere-turns or increase face area | Extra turns add heat, not flux |
| Holds cold, releases hot | Check hot holding force against end-of-stroke spring force | A stronger spring demands more force than it returns |
| Model over-predicts by a consistent factor | Your leakage coefficient or your gap figure is optimistic | Re-tuning the material data hides the geometry error |
| Force varies unit to unit | Check friction and plating thickness before material | Friction scatter usually exceeds magnetic scatter |
When it IS the harder problem
The stroke is long relative to the pole face. Fringing then stops being a correction and becomes the dominant term, leakage grows with gap, and the simple expressions lose their shape entirely. This is where finite element analysis earns its cost.
Part of the magnetic circuit is saturated at the working point. Then force is no longer a clean function of ampere-turns, and the design has a knee that moves with temperature. Build the force-against-current curve and look for where it bends.
The load includes a spring that is compressed at the end of stroke. The requirement is then the difference between two curves that both vary with position, and the worst case is rarely at either end of the stroke.
The supply is AC. Then the gap sets the impedance, the current depends on plunger position, and the whole calculation changes shape. That case is dealt with in AC versus DC solenoids.
A closed-form calculation is worth running even when it is not exact, because it catches the errors that matter most: a gap that is wrong by a factor of two, a leakage coefficient that was never considered, a design with no margin above its own spring. Get the order right, correct for temperature and gap, and then measure once to close the loop.
calculate solenoid forcesolenoid holding force formulaampere turns temperatureair gap force relationshipsolenoid force calculation
Frequently asked
- Which formula should I use to calculate solenoid holding force?
- Either F = B squared times A divided by two mu-nought, or F = (NI) squared times mu-nought times A divided by two g squared. The first is easier to sanity-check because it works from a flux density you can estimate; the second is easier to use from a winding design. Both assume ampere-turns are constant, which is the assumption you have to correct.
- How much does temperature reduce holding force?
- Work it from the copper coefficient. Resistance rises 0.393 % per kelvin, so a 70 K rise above the measurement temperature adds about 27.5 % to the resistance. At constant voltage the current falls to roughly 78 % of its cold value, and because force goes with the square of current it lands near 61 %. A design with less than about 1.6 times margin on paper can therefore be marginal in the real enclosure.
- Why is the air gap in the formula not the air gap on the drawing?
- Because the drawing shows the gap at one nominal condition. Mounting face flatness, plunger perpendicularity, stop wear, plating thickness and the residual gap left by any non-magnetic shim all stack into the assembled figure. The relationship is inverse square while the circuit is unsaturated, so 0.05 mm added to a 0.30 mm gap already costs about 26 % of the force.
- Is a closed-form calculation accurate enough to trust?
- Treat it as good to roughly plus or minus 20 to 30 %. That is enough to catch an order-of-magnitude error and to compare two designs, but not enough to promise a number to a customer. Use it to set the direction, then close the last of the error with one measurement on a real part, and use finite element analysis where the stroke is long or the circuit is partly saturated.