For quite some time, I have been curious about what would happen if I replaced the capacitor in my ceiling fan with an inductor. Here in India, we usually deal with a mains voltage of about 220 - 250VAC RMS at 50Hz.

Here's a diagram of the fan's internals that we often use:

enter image description here Source

The fan originally had a 2.5uF capacitor, which gives it a reactance of 1 /(2 x π x 50 x 2.5e−6) = 1273.88 Ohms.

To match this reactance, I calculated that I would need an inductor of about 1273.88 / (2 x π x 50) = 4.05mH.

I hand wound an inductor that had ~4.05mH@100Hz:


Inductor measurement

I tested it to ensure it could handle the necessary current. The performance graph showed a linear response up to 10 Amps (RMS), well above the 0.25 - 0.3 Amps required by the fan. This indicates that the inductor should not saturate at the fan’s operating current.

After installing the inductors, I found that, as expected due to the 180-degree phase shift compared to the capacitor, the fan began rotating in reverse. To fix this, I swapped the line and capacitor connections, which corrected the rotation direction but led to a new issue: the fan now spins with noticeably lower torque and doesn't produce as much airflow as it did with the capacitor.

Why isn't this setup replicating the fan’s torque as it was with the capacitor? Could this drop in torque be due to harmonics? What's going on here?

  • 1
    \$\begingroup\$ How did you measure the 10 A supposed saturation limit? \$\endgroup\$
    – winny
    Apr 24 at 9:00
  • \$\begingroup\$ Have you checked whether the two motor coils that you swapped are equivalent? Sometimes the capacitor coil has more windings and thinner wire, and passes less current. \$\endgroup\$
    – jpa
    Apr 24 at 13:22

3 Answers 3


Consider not just the equivalent reactance of the components but also the impedance when you have an inductor in series with an inductor and also when you have an inductor in series with a capacitor - the reactances don't just simply arithmetically add.

The capacitor cancels some of the impendence presented by an inductor alone. This, in turn, allows greater current to flow. More current amounts to more torque from the stator. You can check both of your configurations by measuring the current draw with an ammeter.

But that's not all, the substituted inductor will present a reversal in direction as you have seen, but the magnitude of the phase difference will be considerably less with the inductor - an optimized capacitor will bring the stator phases near 90 degree offset: the operating condition to maximize torque (all other factors being equal). The inductor you are using instead may do something like 10 or 15 degrees worth of phase shift. Maybe even less.


You have a single-phase capacitor-start motor (b) or capacitor-run motor (a), which uses the capacitor to create a 90° phase shift so the motor will spin.

Referenced drawing is wrong. There is no coil off of common terminal (C).

enter image description here


You replaced a key component for the motor to work and wonder why the motor has no torque.

Possibly you have created a resistance split phase motor (d).

Single-phase ac when applied to a coil will create an electro-magnet which alternates but does not rotate. The capacitor provides a 90° phase shift from the source voltage, which when applied to a second smaller coil, will create a rotating magnetic field to pull the rotor and create torque.


It's not clear what core you used, but given the inductivity is about 3.54 µH/t2, it sounds like a high permeability ferrite core. Also guessing it's about 2" OD, so is probably around a TX51/32/19-3C90 (AL = 3.98 µH/t2, le = 125 mm, μr = 2300) which will saturate with about 17At (0.4T). Or at 34t, maybe 0.5A, or 0.36A RMS.

It's not clear what impedance the motor winding is, but it's likely much higher than 4mH, and 4mH is only 1.5Ω, whereas the 2.5μF has 1061Ω reactance at 60Hz.

This is a small motor so it might not actually be drawing that much current, but if it's stalled, it might be more, in which case your inductor becomes saturated and is an even lower impedance.

For proper motor operation, you need additional phase shift anyway; perhaps a series resistor acting against the winding's inductance, or a series inductor then a resistor in parallel with the winding. This is rather inefficient of course, so the capacitor (and winding inductance) is preferable.

  • \$\begingroup\$ Yes, I'm using a high permeability ferrite core; you can view the datasheet here: cosmoferrites.com/product-size/toroidal-cores-rings/t4919 . The core material is CF196. Additionally, the frequency for this single-phase fan is 50Hz, not 60Hz, which results in a reactance of approximately 1273Ω. While the fan does rotate, it initially spins in the reverse direction and with reduced torque. Even after reversing the connections, the fan rotates correctly but still with diminished torque. Could this decreased torque be due to a different phase angle than measured by the LCR meter? \$\endgroup\$
    – 15 Volts
    Apr 24 at 13:16
  • \$\begingroup\$ You will have to study in more detail: 1. why a motor rotates, and 2. how to work with phase of currents and voltages in circuit (AC steady state analysis, phasors). In short: the phase measured by the meter, is only one factor among many in the circuit, and you must know all components to solve for phase angle. \$\endgroup\$ Apr 24 at 14:18

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.