Marcus Thielen Provides Insights for Producing Code-Compliant LED Signs

Light-pollution discussions have become a frequent local-government topic in the last few years. This has generated additional requirements for controlling signs’ light levels on sign-permit applications.
The Illumination Engineering Society (IES) and the Intl. Dark-Sky Assn. (IDA) commonly propose such measures (see ST, January 2012, page 34). Dated June 15, 2011, the latest Model Lighting ordinance can be found at www.darksky.org.
Every signshop (and property) owner should read this document carefully, and verify whether these regulations violate Constitutional rights. I believe they have significant societal impact. Illuminated signage’s role in U.S. heritage in special. I especially believe this is illegal in light of Reed vs. Town of Gilbert.
In addition to the tables in the IDA document, which prescribe absolute limits for luminous flux and bright-ness, of interest for signshops are:
C. Lighting Control Requirements
1. Automatic Switching Requirements. Controls shall be provided that automatically extinguish all outdoor lighting when sufficient daylight is available using a control device or system, such as a photoelectric switch, astronomical time switch or equivalent functions from a programmable lighting controller, building automation system or lighting energy-management system, all with a battery or similar backup-power device.
2. Automatic Lighting Reduction Requirements. The Authority shall establish curfew time(s) after which total, outdoor-lighting lumens shall be reduced by at least 30% or extinguished.

Every signshop might check, before filing for a sign permit, if such measures exist locally. I had a case in which the sign was located in a city-business zone, but the other side of the street was considered “residential.” Consequently, the sign had to be completely redesigned due to its light emission.
Local sign codes may require a surface-brightness limit, as well as a total, lumen-output limit (per sign surface or in general), independent of the ambient-light level. There may also be daytime or even observed-holiday restrictions.
The “astronomical time clock” cited in the regulatory text is by no means the well-known time switch. It’s a clock that calculates the sunrise and sunset for each day and location. Such clocks employ some math; the sign’s latitude and longitude must be entered into the software (search the Internet for “astronomical time switch” available models). I also don’t know what “battery or similar backup power or device” implies – the entire sign and control system should be battery operated? Or only so the settings aren’t lost by a mains power failure?
With today’s technology, such control is simple, but it impacts component selection when planning the sign. Some components aren’t compatible with dimming – alone or in combination. As previously discussed (see ST, March 2014, page 34), we must differentiate LED lighting systems in terms of constant current and constant voltage.
Lighting components designed for a series circuit, operating on constant current, need explicitly dimmable power supplies. They can’t be dimmed by reducing the operating voltage, the phase angle of the mains power, or the switching (Pulse Width Modulation [PWM]) principle.
Parallel-wired (constant-voltage operation) LED modules are available in two varieties – with the resistor (passive) current stabilized, and the active, regulated (with linear semiconductor control) ones. The former can be dimmed by reducing the operating voltage or PWM; the latter can be dimmed only by PWM.
For high-power, LED modules – used mainly in cabinet signs – another circuit variety has recently claimed some marketshare: the active switching regulator (Fig. 1). It can’t be dimmed at all, unless the regulator on the module has a special design and dimming-control input (I’ve not seen this in my laboratory). Note that high-voltage LED modules (those connected directly to the mains power without an additional power supply, like the Seoul Semiconductor Acriche series) aren’t dimmable at all.
Traditionally, illuminated signs were controlled either by a clock switch, a daylight sensor or a combination of them. This permitted the sign to be on during darkness, or during the evening and early morning (off at late night), or a combination of both conditions.
Also, “traditional illuminated signs” were designed so, even when not illuminated during the day, the sign would be readable and transmit its message (Fig. 2). But modern, LED-type signs, or backlit digital prints, often aren’t visible or can’t transmit their message when not illuminated, so electrical illumination is always needed.
Meanwhile, the neighborhood – even in industrial or commercial zones, not only in the city centers – claims its right to a good night’s sleep. With animated signs or large video screens with changing content, illuminated-signage annoyance can be a problem. Additional restrictions may apply in such situations.
Thus, sign codes often require variable light intensity for day/nighttime and daylight levels. This requires “remote” dimming – not a manual, intensity-control knob you set once. And the light controller needs to have more “intelligence” than a simple time clock or daylight sensor; hence, the “astronomical clock.” Sometimes, the brightness of the sign itself must be measured, and you must employ closed-loop control of the brightness level.
However, digital, LED videoscreens don’t have a “brightness” control knob like TV sets. Unlike lightbulbs, dimming by reducing the supplied power doesn’t work either.LED videoscreens usually combine tiles; each tile is a separate unit. Each unit’s LEDs vary due to the manufacturing process. A screen needs to be “calibrated” after assembly so all tiles give “the same” white and intensity (which in practice differs and creates visible flaws of such screens).
Reducing total brightness can be achieved by either altering the calibrating parameters – or by a program that actually gives the tiles their image data (a quasi RIP program to match the video resolution to the screen pixels and separate the partial images according to the tile sizes and positions). Thus, dimming an LED screen always involves changes in the control computer’s operating program. Here, a “patch” or additional routine must be implemented to reduce the LED brightness the sign-code restrictions address.
If the calibration data is maintained, each pixel usually gets an RGB value for the partial intensity in the composite colors red, green and blue. With eight bits per color, 255 is the maximum intensity. Reduction of total brightness means each pixel in every image needs three multiplication operations (an eventual floating point).
For example, consider a resolution of 1024 x 768 pixels. At 25 images per second, with three composite colors/pixels – you’d need 58,982,400 floating-point multiplications every second, just for correcting the brightness/RGB values!
Because LED-screen brightness control always involves control software, check with the manufacturer to confirm what is used, and the additional cost for a time/ambient, light-level control. Adding this after a sign is installed is often problematic, if not impossible (not enough computer power, or input, etc.)
Most current LED screens have Internet access for updating media content. Thus, the computer’s clock can be the timesource for the dimming requirements, and content can updated from a time server (NTP) via the Internet.
For “simple” signs, even a quartz clock can and will drift in time (what about daylight savings time?). I recommend using precise time information from a GPS receiver. Note: This time information is usually given in UTC (universal time coordinated), and local (daylight savings) time must be calculated from the date info by the control computer. Consequently, not every LED sign system is dimmable, as local regulations might require. Thus, a check of the local sign ordinance requirements prior to quoting a new sign might be a good idea.