The news about lighting is that there has been a revolution in solid-state lighting (SSL).

The extent of the revolution can be gleaned from Dr. Mark Rea, director of the Lighting Research Center (LRC) at Rensselaer Polytechnic Institute in New York. “There was a story line in lighting that you never displace a light source, you can only add a new one to the family,” he says. “LEDs have changed all that.”

Rea says that LEDs can compete with any lighting technology and do a better job. Because LEDs use so little power, the only cost-effective energy-saving lighting control strategy involves an occupancy sensor and a switch. Rea says that the old “ballistic” model of control that was used until a few years ago makes less economic sense. That model is based on the notion of launching an on/off or dimming message and seldom if ever looking for feedback.

“Now it’s two-way communication,” he says. “Controls are very different philosophically than they were even five years ago. In the future, the data acquired with new controls may be even more valuable than the energy those controls save.”

Mark Rea, RPIMark Rea, RPIThe minimum requirements for a two-way LED lighting control strategy are occupancy sensors, switches, dimmers, and networks.

Occupancy Sensors

The most common occupancy sensors use one of two technologies, or a combination of both.

Passive Infrared (PIR) senses the difference between the heat emitted by humans and background heat. A space is assumed to be occupied when the human heat is in motion. PIR requires a direct line of sight from the sensor to the occupant and is effective at detecting people walking into or out of a space. However, it only works for sensing major motion.

(Learn more about light emitting diodes through Engineering360 product and supplier guides.)

An ultrasonic sensor emits a signal that is reflected off all objects in an area. It detects motion by the Doppler effect, namely, the shift in frequency between the emitted and reflected signals if an object is in motion. Ultrasonic sensing can cover a space even if obstacles are present; it also can detect small motions. The problem is that it can be sensitive to motions caused by vibrations or air currents.

Dual technology uses both methods in combination for more reliable operation. In order to turn lights on automatically, both sensors must detect someone entering a room. After that, only one of the two is needed to keep the lights on while the area is occupied.

All three versions are available in ceiling-mount or wall-mount configurations.

Some energy codes such as ASHRAE 90.1 and New York City’s Energy Code LL48 mandate that lights be switched on manually when the first person enters a space and switched off automatically when the space is no longer occupied. However, it is sometimes acceptable to switch the lights on automatically to 50% of the normal illumination level when a person enters, requiring a manual switch for full illumination.

Using sensors to turn off lights or dim them when a space is unoccupied can save energy. Even in offices that are generally used all day, workers often arrive and leave at different times, rendering timer control counterproductive.

Different spaces, such as open plan offices, partitioned offices, private offices, utility rooms, cafeterias, conference rooms, hallways, and lobbies have different lighting needs.

A selection of commercially available LED lamps with Edison (screw-type) base. Image source: wikipedia, credit Geoffrey A. LandisA selection of commercially available LED lamps with Edison (screw-type) base. Image source: wikipedia, credit Geoffrey A. LandisIn any particular application, it is necessary to know the exact specifications of a sensor to decide how it should be mounted for coverage of the targeted space. For example, in an open office with cubicles, ceiling-mounted sensors should probably be specified and arranged so that their coverage patterns overlap. (Coverage patterns are specified by the manufacturer for each device.) In such an office, it often is a good idea to choose a longer time for automatic off, say 15 to 30 minutes.


There are a variety of methods for LED dimming. The following basic facts should be considered in their design and selection.

· Their light output is directly proportional to current.

· The relationship between LED voltage and current is nonlinear.

· Correlated Color Temperature (CCT) of an LED is specified at a given voltage/current operating point.

· Efficacy in lumens/watt is specified at a given operating point.

· Switching times between emitting and non-emitting states are on the order of magnitude of nanoseconds.

· LEDs have a significant startup inrush current — the peak to average current ratios can be as high as 30:1.

LEDs require drivers to convert AC line voltage to low voltage DC because LEDs are semiconductor diodes. An LED dimmer must be designed to be compatible with the driver and the driver with the specific LED.

When replacing an incandescent lamp with an LED, it is important to change the dimmer as well. Typical incandescent dimmers use so-called phase-cut technology, where the conduction angle of each half cycle of the input sine wave is varied to control average power. Attempting to use them with LEDs can lead to unstable performance.

Two common dimming technologies exist for LEDs: pulse-width modulation (PWM) and constant current reduction (CCR).

In PWM, the LED current is switched between zero and rated output. The ratio of on to off time is varied to control average luminous intensity. For example if on time is equal to off time then the intensity is 50% of maximum. The switching is done at a high enough frequency to avoid detectable flicker.

The advantage of PWM is that the LED is always switched on to rated current so color temperature and efficacy are constant across the dimming range.

One disadvantage of PWM is that the fast rise and fall times of the high frequency on/off pulses can generate electromagnetic interference (EMI). Performance issues also may arise if there are long wire runs between the dimmer and the light source. What’s more, the power source for the LED must be rated to handle the peak turn-on transients, not just average power.

CCR is an analog technique in which the voltage is held constant and the DC current is varied to control luminous intensity. It avoids the problems due to high frequency switching, but CCT and efficacy will vary with dimming percentage.

Retrofitting Lighting Controls

The challenge of how best to retrofit existing lighting and controls is a hot topic. For simple on/off control, either from a manual switch or an occupancy sensor, there’s no problem — simply replace the lamp. Screw-base lamp replacements typically have self-contained drivers. For fluorescents, however, the ballast must be removed from the fixture and replaced with an LED driver.

To reap the benefits of LED lighting systems, a network for communication and control needs to exist. Perhaps the most basic challenge for establishing a network is how to physically interconnect the lamps and controls. This can be more difficult for retrofits than for new construction.

Commercial buildings can benefit from LED technologies that support energy management. Image source: wikipediaCommercial buildings can benefit from LED technologies that support energy management. Image source: wikipediaDimming is a good example. Because existing incandescent and fluorescent dimmers won’t do the job, the control function must be separate from the main power source. Although different approaches exist to solving this problem, they fall into two general categories: wired and wireless.

Power over Ethernet (PoE) has many advantages among the wired systems. PoE luminaires (code-talk for lighting fixtures) typically include drivers, occupancy sensors, and dimmers.

The system uses standard Ethernet category style cable to power the LEDs and also to transmit bi-directional control and sensor signals. Cat 5 cable can deliver up to 51 watts at 37 to 57 VDC. An upgrade to the IEEE 802.3 Standard is expected to be approved in 2017, which will increase the source power to 90 watts. The most efficient LED lighting fixtures for offices draw 30 – 35 watts.

The beauty of PoE is that it enables each luminaire to be configured as an Ethernet network node with a unique IP address so that a control system can interact with each luminaire individually. This opens a world of possibilities, especially in terms of the data acquired from the installed lighting.

The control and sensing information can be handled with a dedicated controller or with software that can be uploaded to a computer or mobile device. Sensor data, such as on/off/dimmed status, occupancy, and temperature can be shared via Ethernet with Building Automation Systems (BAS) and Enterprise Resource Planning (ERP) systems. It can also be shared with other enterprises via Internet, and may be sent to the cloud for processing, storage, and analytics. The system software also can be reconfigured for future upgrades or expansion as requirements expand and the technology evolves.

One issue with PoE, however, is that it requires running cable to each luminaire. The wiring does not have to be in conduit so it can be easily installed above a drop ceiling and typically does not require a licensed electrician. However, the work should be completed by someone certified for low-voltage wiring.

A wireless system is a much less costly option for retrofitting — one must invest in some hardware but not in labor-intensive wiring.

A number of wireless technologies are on the market. Some, such as Lutron’s, are proprietary, and can be used only with Lutron products. Others use standard protocols such as Z-Wave and ZigBee; a new one designed for home automation is called Thread.

One choice that might seem surprising is Bluetooth. Originally started in 1994, it is thought of primarily as a short-range cable-replacement technology. That is changing, however.

The next generation of Bluetooth, commonly known as Bluetooth Smart, was introduced in 2010 and may offer advantages compared to the other wireless technologies.

For one thing, it can transfer data at 1 Mbit/s, which is orders of magnitude faster than Z-Wave and Zigbee. There are advantages to the high data rate even though building automation does not require anything like that throughput.

It also enables lower latency, and thus better responsiveness, and also enables lower-duty cycle for transmitting information. That, and support for sleepy nodes, which only wake up when polled, extends battery life for sensors. Frequency hopping--the ability to automatically select the clearest among 40 different channels--is another possible advantage. A Bluetooth upgrade, to be released in the first half of 2017, is expected to double the speed and multiply the range by four.

A drawback of Bluetooth, however, is that the core specification does not support mesh networking. However, a number of companies have developed proprietary methods for using Bluetooth with mesh networks. And the Bluetooth Special Interest Group (SIG) is working on incorporating a standardized method for that.

(See "A Tale of Five Protocols" for a comparison of wireless protocols.)

Load Shedding

One important way that a lighting control system can contribute to improving energy supply is through load shedding. Indeed, utilities have a problem: although they can plan for average demand, there are unpredictable spikes. This might occur, for example, on an unusually hot summer day. The utility must therefore provide enough capacity to handle the peak loads even if they rarely occur.

One method of satisfying a spurt in demand is with “spinning reserve,” which is the ability of a power plant already connected to the power system to increase its output. The other option is non-spinning reserve — increasing output by bringing fast-start auxiliary generators on line or importing power from other systems.

However, if the demand spike is especially high, the response can generate instability on the grid. In any event, dealing with demand spikes is expensive for utilities. So, in addition to charging commercial customers for total consumption, utilities often add a surcharge for peak demand, typically the highest load for any 15-minute period during the billing cycle.

One solution may be to reduce spikes in demand. Networked LEDs can play a role in that, especially in control systems for microgrids. The grid can send a signal to its customers that their load should be reduced, so-called load shedding.

As Rea says, "you can't dim your copier or your computer but you can dim your lighting" if it’s part of a network.

A study published in the Lighting Research Technology Journal [Lighting Res. Technol., 37, 2 (2005) pp. 133 – 153] investigated the effects of reduced light levels on office workers. The researchers concluded that a reduction in luminance of 15% for paper tasks or 20% for computer tasks was undetectable for half of the participants in their study.

Further, most employees would accept dimming of 30% for paper tasks and 40% for computer tasks even though these levels are detectable. Notably, if the importance of load shedding is explained to employees, the acceptable dimming limits could be increased to as much as 40% for paper tasks and 50% for computer tasks.

Since the power used by LED lamps is approximately proportional to their illumination level, a 50% dimming of lighting would result in a 50% reduction in power demand.