An inductive appliance or equipment (fan, magnetic ballast etc) can sometimes cause interference with a dimmer when it is installed in close electrical proximity to the dimmer.  The lights on a dimmer might exhibit a shudder, flicker or a quick flick and in may turn off when for example an older magnetic ballast or a fanis turned off (or changing speed).  This usually occurs when the inductive equipment, the switch or speed controller is installed close (electrically) to the dimmer as per the diagram below.

This is due to electrical interference causedby mainly the turn off event of the inductive equipment which typically follows the path of least electrical resistance.  In this case entering the dimming circuit. The interference is however not limited to dimmers, it can cause lines across an older analogue TV or cause noise on a radio, etc.  This phenomenon does not occur every time that the inductive equipment is turned off and is thus sometimes perceived as a random fault. The interference and “randomness” thereof are due to:

  1. An inductive load stores magnetic energy and when it turns off may produce a voltage spike or back EMF.  This voltage pulse may trigger the dimmer’s internal over-voltage protection or it may result in an overcurrent into the dimmer load, triggering the dimmer’s overload protection if the dimmer is equipped with these protection mechanisms. Either case will turn the dimmer off during the fault event and depending on the dimmer design, it might turn on automatically at the next AC cycle or remain off until manually turned on again.  The back EMF is largest when the current through the inductive equipment is the largest and depends on the power factor of the equipment. If it has a pure inductive power factor, then the back EMF is the largest when turning the equipment off at the AC zero crossing.  Since the switch might be turned off at any amplitude of the voltage and in cases at amplitudes where the back EMF is small, the resulting dimmer reaction seems random.
  2. The above voltage pulse can result in arching of the switch and can sometimes be observed as an audible arching or sparking noise in the switch.  This generates severe electrical noise which may last for several milliseconds or even extend across AC cycles.  This noise can also trigger the dimmer’s internal protection but often causes the dimmer to not sense the AC zero crossing accurately, causing it to skip a cycle or portion thereof, resulting in flickering or a flash of the light. Since the arching is largest at the highest back EMF and since the interference might occur at a position in the AC cycle where the dimmer is not sensing the mains zero-crossing, the perceived “randomness” of the dimmer interference increases.

To demonstrate the interference, an inductive “wire-wound” ballast with a mechanical switch was wired to the supply of a dimming circuit with a single LED load as per the above diagram and the load voltage and current measured.  In all cases,the red waveform in the oscillograms is the lamp voltage at 100V/div and the yellow trace the lamp and dimmer current at 50mA/div.  The time scale is 5ms/div.

The ballast was turned on and off repetitively until a visible flick or flicker was observed in the lamp on the dimmer and the relevant waveforms captured at the instance of the interference.  Thesame tests in the same test setup were conducted with the following trailing edge LED dimmers:

  • Diginet 400W rotary dimmer (Australian)
  • Intellibus/R&D 350W “Beta 1” bell-press dimmer
  • Intellibus/R&D 350W “Beta 2” bell-press dimmer
  • Shuttle 500W bell-press dimmer

The oscillogram shows the reaction of the Diginet dimmer:  it can be seen that directly after the severe interference, the dimmer fails to recognise the next AC zero crossing and skips a cycle, resulting in a lamp flicker.

The same behaviour can be observed with the 2 Intellibus/R&D dimmer models: after the interference that overlaps the AC zero crossing, the dimmers fail to register the next AC zero crossing and misses an AC cycle and the lamp flickers.

During the measurement with the Shuttle dimmer a double interference was captured which shows that the dimmer reacted the same as the above dimmers for the first interference instance where it failed to register the next AC zero crossing, but the following interference did not cause a reaction and the dimmer operated correctly during the seemingly worse interference condition.

This is becausethe interference is a series of extremely quick pulses which, at the capturing scale looks like a much lower frequency.  The conditions were thus such during the second set of interference pulses that the dimmer did not experience interference at the instance that the AC mains voltage crossed from positive to negative.

This is a good example which demonstrates that the dimmer will not always react to the interference and below is an example with the Shuttle dimmer where the dimmer and lamp remained perfectly stable during the inductive interference which occurred well before the AC zero crossing.

When wiring an“EMI” capacitor across the inductive equipment or the complete inductive circuit as per below diagram, nearly no electrical noise was measured,and all the dimmers remained perfectly stable during repetitive switching of the inductive circuit.  This is due to the fundamental operation of a capacitor which is a low impedance to a high-frequency pulse (Zc = 1/(2xfxC)) and conceptually circulates the noise in the inductive/capacitive circuit, preventing it from emitting or conducting to external circuits.

The capacitor must, however, be an X2 type (double insulated) and rated at 275VAC or higher.  A 0.22uF (220nF) value capacitor is usually a good choice, but in cases of severe interference,a higher value such as 0.33uF or 0.47uF can be used.  This is incidentally the same part and value that is connected to the live and neutral of hairdryers, drills, etc. to prevent the electrical noise from the brushes from interfering with radios and televisions.