Most of the trouble I have seen with three phase motors happens in the first two seconds after the contactor closes. The motor either gets up to speed or it does not, and whatever damage is being done to the windings, the coupling and the supply is done in that short window. It is worth understanding what is physically happening in there rather than treating starting as a black box that either works or trips the breaker.
The short version is that an induction motor is a rotating transformer with a short circuited secondary, and at the instant of starting that secondary is standing still. Everything else follows from that.
The rotating field comes first
Three stator windings are spaced 120 electrical degrees apart around the bore. Feed them with a balanced three phase supply and the combined magnetic field does not pulse in place, it rotates. The speed of that rotation is the synchronous speed, and it is fixed by the supply frequency and the number of poles. At 50 Hz a two pole machine gives 3000 rpm, a four pole gives 1500 rpm, a six pole gives 1000 rpm. At 60 Hz the same machines give 3600, 1800 and 1200 rpm.
The rotor is not connected to anything electrically. It is a stack of laminations with aluminium or copper bars shorted together by end rings, which is why it is called a squirrel cage. When the stator field sweeps past the stationary bars it induces a voltage in them, current flows because the end rings close the circuit, and the interaction of that rotor current with the stator field produces torque. The rotor accelerates in the direction the field is turning.
The rotor can never quite catch the field. If it did, there would be no relative motion, no induced voltage, no rotor current and no torque. The difference is slip, usually 1 to 5 percent at full load on a general purpose machine. A 1500 rpm nominal motor running at 1455 rpm is sitting at 3 percent slip and that is perfectly normal.
Why the starting current is so high
At the instant of starting, slip is 1.0. The rotor is stationary and the field is sweeping past the bars at the full supply frequency. The induced rotor voltage is at its maximum and the only thing limiting rotor current is the impedance of the rotor circuit itself. Referred back through the magnetic coupling, the motor looks to the supply roughly like a short circuited transformer.
That is why locked rotor current on a direct on line start is typically six to eight times full load current, and on some high efficiency designs it reaches nine or ten times. The consequences of that current are the ones you actually deal with in a plant.
- Voltage dip on the supply, which can drop out contactors elsewhere and make lighting flicker visibly.
- Heating in the rotor bars and the stator winding, concentrated and fast, because there is almost no time for the heat to go anywhere.
- Mechanical shock through the coupling, the gearbox and the driven machine, since torque appears in a few electrical cycles.
- Nuisance tripping of upstream protection that was selected on running current rather than starting current.
There is also a transient on top of all this. Depending on where in the voltage waveform the contacts close, the first current peak can sit well above the symmetrical locked rotor value, and the brief torque spike that goes with it is what shears keys on badly fitted couplings.
Reading the torque speed curve
The useful picture is torque plotted against speed from standstill to synchronous. Four points on that curve have names that appear in specifications and you should know all of them. Locked rotor torque is what the motor produces at zero speed. Pull up torque is the minimum value somewhere in the low to middle speed range, and it is the point where a motor most often stalls against a stubborn load. Breakdown torque is the peak, usually between 75 and 90 percent of synchronous speed. Full load torque is where the machine settles with its rated load applied.
The shape of that curve is a design choice. NEMA MG 1 defines design letters for it and IEC 60034-12 defines starting characteristic classes such as N and H. The differences come mostly from rotor bar geometry, because deep bars and double cage rotors use skin effect to raise rotor resistance at the high rotor frequency present during starting, then let it fall as the rotor comes up to speed.
| Design | Locked rotor torque | Starting current | Slip at full load | Typical use |
|---|---|---|---|---|
| NEMA A | Normal to high | High | Under 5 percent | Machine tools, injection moulding |
| NEMA B | Normal | Normal | Under 5 percent | Fans, pumps, general purpose |
| NEMA C | High | Normal | Under 5 percent | Loaded conveyors, compressors |
| NEMA D | Very high | Low to normal | 5 to 13 percent | Punch presses, hoists, high inertia |
The practical point is that a design B motor substituted into a design C application will start, badly, on a good day. On a cold morning with stiff grease and a full conveyor it will sit near pull up torque drawing locked rotor current until something upstream gets tired of it.
Inertia sets the starting time
Acceleration time depends on the difference between motor torque and load torque across the whole speed range, divided into the inertia that has to be accelerated. Take a 22 kW four pole motor at 1470 rpm. Rated torque is roughly 9550 multiplied by 22 and divided by 1470, which is about 143 Nm. If the average accelerating torque over the run up is 1.5 times rated, that is about 214 Nm available, and if the total referred inertia is 4 kg m squared then the angular acceleration is about 53 rad per second squared. Getting to 154 rad per second takes roughly 3 seconds.
Three seconds is fine. A large centrifugal fan with 60 kg m squared of inertia on the same motor is a different conversation, because the start stretches to half a minute and the rotor cage is absorbing serious energy the whole time. That is when you start checking the motor thermal withstand time and the safe stall time, and when the manufacturer starts asking how many starts per hour you intend to do. Two cold starts and one hot start per hour is a common limit on larger frames.
What actually goes wrong on site
Single phasing during the start is the classic killer. Lose a phase with the motor stationary and it will not turn at all, it will simply sit there humming and drawing heavy current through two windings until the overload responds. Lose a phase while running and the motor usually keeps going at reduced capability, which is more insidious because the current imbalance cooks the windings over weeks.
Undervoltage is the other common one, and it bites harder than people expect. Induction motor torque varies with the square of the applied voltage, so a 10 percent dip leaves only about 81 percent of the torque available. If the motor was already marginal on a high inertia load, that is the difference between a slow start and a stall.
Finally, check what the driven machine is doing. Half the stalled starts I have investigated were mechanical. A dragging brake, a seized idler, a pump full of settled solids, a gearbox filled with the wrong oil in winter. The motor is the thing that trips, so it gets blamed, but it is usually the honest party. Background on the machine itself is well covered in the Wikipedia article on induction motors, and the governing performance standards sit under the IEC 60034 series published by the International Electrotechnical Commission.