In commercial lighting applications with HiD or older 2 pin fluorescent lamps, a magnetic ballast is used to control the energy in the arc lamp. In the first generation of LED replacement lamps, two kinds were available:
- LED lamps that only worked with the ballast – such as Lunera’s Gen1 & Gen2 Susan families – known by DLC as a “type-A” lamp;
- LED Lamps that only worked when the ballast was removed – such as the many corncob lamps available from China – known by DLC as a “type-B” lamp.
However, more recently Lunera has introduced universal input products that support both type-A and type-B operation. Many people ask us how the lamp knows which mode is running it – whether it is talking to a ballast or directly to line voltage.
How a Magnetic Ballast Works
There are broadly two kinds of magnetic ballasts: 1) reactor ballasts and 2) autotransformers.
Reactor ballasts are the simplest – they use a large inductor (or coil) in series with the lamp and, sometimes, a capacitor to help clean up the power factor of the resulting load (fig. 1). The coil as a specific impedance at 60 Hz and an arc lamp operates at a relatively fixed voltage. Thus, the current is simply:
Fig. 1 A reactor ballast, which may or may not include a capacitor (used for power factor compensation). A reactor ballast can only be designed for one input voltage (i.e., only 120V or 277V, not both).
When engineering an LED lamp, a type-A lamp typically uses the coil in the same way that the legacy lamp did – to control the output current. Power is simply Lamp current * Voltage. It can be calculated for a set input parameter and the desired operating voltage of the lamp configured and the resulting circuit current calculated.
The downside of the type-A design is two-fold:
- Users often have no idea what kind of ballast they have installed in the infrastructure. It causes a significant number of application-related failures when the lamp is installed in an application it is not suited for.
- The resulting power factor of the combined load is low – either because there is no compensation capacitor in a non-compensated or because the compensation capacitor is too large for the LED load in a compensated ballast.
An autotransformer ballast is slightly more complex, but the net result in similar. The purpose of the autotransformer is to allow for multiple input voltages on different input taps and transform that voltage into something suitable for the lamp, while minimizing the magnitude of the required reactor section (fig. 2).
Fig. 2 An autotransformer ballast with 4 input taps. As in the case of the reactor ballast, this transforms the constant voltage from the utility to a constant current to the lamp; however is more flexible and more tolerant to variances in line voltage.
A type-A LED lamp can be designed for an autotransformer in the same way as a simple reactor ballast, and suffers from the same disadvantages. It will generally not achieve high power factor or low THD and is compatible with one or a small family of ballast types.
Type-B Lamps Operation
A type-B lamp uses a switching power supply in a variety of architectures to convert an input AC voltage (120V, 277V, etc.) to the output voltage needed by the LED array. These lamps have been around for some time (fig. 3).
Fig. 3 An autotransformer ballast with 4 input taps. As with the reactor ballast, this transforms the constant voltage from the utility to a constant current to the lamp; however is more flexible and more tolerant to variances in line voltage.
A type-B lamp switches on and off at high frequency (typically 50-250kHz) to let current in or not; then uses a variety of energy storage and filtering devices (inductors and capacitors) to transform this switching behavior into a load that looks linear. The lamp will thus pull a constant amount of power from its input over a wide input voltage range.
Universal Magnetic Ballast Type-AB operation
With careful design and characterization, it is possible to engineer a universal input lamp. It is a type-B lamp that works well with a whole variety of reactor and autotransformer ballasts as well as line voltage, which is what a type-AB lamp does. The design challenges of doing so reliably are considerable (fig. 4):
Fig 4 – Lineartech LT3799 Flyback LED Driver Reference Design – Ballast vs. Line Performance. In the line case, peak input voltage is 560V and peak LED current is 5.5A. The same circuit driven by a ballast shows a peak voltage of over 1000V and a peak LED current of 7A. This design would likely not function reliably in a ballast driven application.
- There is a positive feedback loop between the reactance of the ballast and the switching power supply which inherently challenges the stability of this architecture.
- The reactance results in a lag between current demand and current flow in a switching power supply
- At the beginning of a half-cycle, the lamp requires a lot of current and appears as a low impedance to the ballast.
- Once the lamp has received its desired current, it asks for much less near the middle of the half cycle; this causes the output voltage of the ballast to increase rapidly, which the lamp must tolerate.
- If not damped appropriately, this oscillation has the potential to create a variety of problems including over-voltage on the lamp input, audible harmonics on the ballast or visible flicker on the lamp.
- It is thus critical to carefully characterize the full family of ballasts to be supported and ensure that these behaviors are managed adequately.
Once done properly, the result can be magical. For example, in our Lunera HID LED 360 8k lumen lamp the performance is:
By selecting an appropriate ballast operating point where the secondary current was low, ballast losses were reduced from 31.3W in a type-A lamp to 12W in a type-AB lamp. It significantly increases ballast lifespan and reduced energy consumption.