Works cold, fails hot: diagnosing thermal expansion in solenoids
A unit that actuates cold and fails hot is almost always running out of current, not out of clearance. Copper resistance rises about 0.393% per kelvin, so a winding at 90 °C has roughly 27% more resistance than at 20 °C and at fixed voltage carries about 22% less current. Measure hot and cold resistance — the ratio gives you the winding temperature and the size of the loss in one step.
Why this happens
Two mechanisms make a solenoid behave differently hot than cold, and they are frequently confused. Sorting which one you have takes one measurement.
The first is electrical. Copper resistance has a positive temperature coefficient of about 0.00393 per kelvin, referred to 20 °C. A winding that reaches 90 °C has a resistance roughly 27% above its cold value. If the supply holds voltage constant — which most do — the current falls to about 78% of the cold figure. Since force in the pre-saturation region rises with current, the available force falls with it. A design that measured 30% margin on a bench at 22 °C can be below requirement at operating temperature without anything being defective.
The second is geometric. Different materials expand by different amounts, so a hot unit can have a slightly different air gap, a different plunger-to-bore clearance, or a different preload than the cold one it was measured as. Force in the working region falls with roughly the inverse square of the gap, so even a few hundredths of a millimetre matters. But this mechanism usually produces smaller numbers than the electrical one, which is why it should be checked second rather than first.
There is a third possibility worth naming, because it does not point to the coil at all: your power supply may itself be drifting. A supply that sags as it warms produces the same symptom from outside the solenoid.
Check these in order
1. Cold and hot winding resistance. Measure with the unit at a known temperature, ideally four-wire. Record resistance and temperature together. Then run to thermal equilibrium and measure again. Rise in kelvin = (R_hot / R_cold − 1) / 0.00393. A ratio of 1.27 means a rise of about 69 K, which in a 25 °C environment means a winding at roughly 94 °C. You now have the temperature, without a thermal model.
2. Voltage at the coil terminals, hot and cold. Not at the supply. If the coil voltage drops when hot, some of your loss is in the wiring, the connector, or the supply rather than the winding. Compare the two readings: if the hot drop is larger than copper alone explains, the supply is participating.
3. Force or stroke, measured hot, at the same point in travel. Measure at the position where the unit actually fails, not at the hold position. The end of the stroke is where force is lowest and where a small deficit becomes visible.
4. Separate “less force” from “more friction”. A hot unit with thinner lubricant should move more freely, not less. If hand travel gets harder as the unit warms, you are looking at clearance closing up or at a part that distorts when hot, and the electrical explanation is not your primary problem. This test takes two minutes and eliminates half the candidate list.
5. Check whether the load is changing, not the driver. Fluid viscosity, a mating part, a housing that clamps down on the assembly as it warms. If the solenoid’s own output is unchanged when measured in isolation, look at what it is pushing against.
6. Only then consider thermal expansion inside the magnetic circuit. Compare the measured gap hot and cold at four points around the plunger. If the gap changes and the current does not, geometry is your mechanism and the fix is a material pairing or a clearance change, not a coil change.
What actually to change
| Finding | What to change | Why not the other thing |
|---|---|---|
| Current loss from copper heating | Raise supply voltage, or drive constant current | Adding turns raises resistance and inductance, making it worse |
| Long hold periods causing the heat | PWM hold or mechanical hold | Removes the heat source rather than fighting it |
| Supply sags as it warms | Replace or derate the supply, upsize the run | A larger coil still sees the sag |
| Gap changes with temperature | Match expansion of plunger and core, or increase clearance | A different core grade does not address geometry |
| Clearance closes when hot | Increase clearance, change material pairing, check coating | More force will not fix a rub that keeps growing |
| Fails only at cold start | Address viscosity or preheat — opposite mechanism | Nothing about the hot-state analysis applies |
When it IS the harder problem
Two mechanisms that both move in the same direction. This is the case that costs the most time, because fixing either one alone does not clear the complaint. Resistance rise takes about 22% of the current at 90 °C; a mounting face that is not quite flat takes a little of the gap; the enclosure adds 25 K of ambient that nobody counted. Each is defensible on its own and none is out of specification. Together they consume the margin, and the unit that was comfortable at 22 °C on the bench is not comfortable in the machine in July. The way out is to measure all three and budget them together, rather than fixing whichever one is easiest to reach.
A measurement taken before equilibrium. Most test benches run for half an hour. Small coils on a metal frame can settle in that time; a potted coil in a plastic housing with a poor conduction path can still be climbing after two hours. If your hot measurement was taken early, your calculated margin is optimistic and you will design to a number that does not exist. Log resistance every 15 minutes and only trust a reading once three in a row agree.
The unit that recovers. When a unit is worse hot and then better after more running, resistance cannot be the cause, because resistance only rises. That pattern means something is loosening or thinning — lubricant distributing itself, a part relaxing, a fluid warming up. I have chased one of these for a week assuming it was thermal derating, when it was a grease that needed a few minutes of shearing before the plunger moved freely. The direction of the symptom, not just its presence, is the clue.
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 fails when hotsolenoid works cold fails hotcopper resistance temperature coefficientsolenoid force drop hotsolenoid thermal effect
Frequently asked
- Why does a solenoid lose force as it warms up?
- Because copper resistance rises with temperature while your supply voltage does not. From 20 °C to 90 °C the winding resistance increases about 27%, which at fixed voltage cuts the current to roughly 78% of its cold value. In the pre-saturation region force tracks current, so a design with 30% cold margin can be under water at operating temperature.
- How do I find the winding temperature without a thermal simulation?
- Measure resistance cold at a known temperature, then again after thermal equilibrium, and use the ratio. Resistance ratio minus one, divided by 0.00393, gives the mean rise in kelvin. A coil whose resistance grows 27% has risen about 69 K. This is the resistance method the insulation standards are written around, and it is more reliable than any surface reading.
- Could it be thermal expansion of the gap rather than resistance?
- It can be, but check resistance first because the magnitude is usually larger and the measurement is easier. If you measure a 22% current loss and the failure is a 20% force deficit, resistance already explains it. Expansion becomes the prime suspect when hot and cold current are the same but the unit still behaves differently.
- The failure resolves itself after the machine has run for a while. Does that change the diagnosis?
- Yes, and it points the other way. A unit that is worse when hot and better after more running is unlikely to be a resistance problem, because resistance only increases with heat. Look at mechanical clearance closing up, lubricant behaviour, or a fluid that thins out as it warms. Both directions of failure exist and they lead to opposite fixes.