WO2006033966A2 - Energy savings device and method for a resistive and/or an inductive load and/or a capacitive load - Google Patents

Abstract

An energy savings device for an inductive, a resistive or a capacitive load, such as a fluorescent light fixture having a magnetic ballast or an electronic ballast, which is powered by an Ac voltage waveform. The energy savings device includes a setting unit for setting a desired power operating level for the load. The energy savings device also includes a processor configured to receive a signal from the setting unit indicative of the desired power operating level for the load, to determine a phase delay to be provided to an output AC voltage waveform that is to be provided to the load, and to output a control signal as a result thereof. The energy savings device further includes an active element provided between a line that provides the input Ac voltage waveform and the load, the active element receiving the control signal and turning off and on at predetermined times in accordance with the control signal, so as to create the output AC voltage waveform from the AC voltage waveform. The processor includes a synchronization circuit that synchronizes to the Green Safety ground line.

Description

ENERGY SAVINGS DEVICE AND METHOD FOR A RESISTIVE AND/OR AN INDUCTIVE LOAD AND/OR A CAPACITIVE LOAD

RELATED APPLICATIONS

[0001] This application is a continuation-in-part of U.S. patent application serial number 10/454,840, filed June 5, 2003, which is itself a continuation-in-part of U.S. patent application serial number 10/205,031, filed July 26, 2002 (which matured into U.S. Patent No. 6,724,157, issued April 20, 2004), whereby this application also claims priority to provisional patent application 60/336,222, filed November 14, 2001. BACKGROUND OF THE INVENTION

A. FIELD OF THE INVENTION

[ooo2] The invention relates to an energy savings device or method that can be applied to a resistive, an inductive, or a capacitive load regardless of the respective impedance or inductance or capacitance of the load. More particularly, the invention relates to a load dimming device that is mounted in series with a resistive, an inductive or a capacitive load and that has access for power and operation to one side of an electrical line supplied to the load. A fluorescent light fixture or a motor for a fan or other device, for example, can be controlled by way of an energy savings device or method according to the invention. The load dimming device can be used to control Compact Fluorescent Lamps, as well as other types of light fixtures, or a motor for a fan or other device.

B. DESCRIPTION OF THE RELATED ART

[ooo3] The ability to control illumination levels is strongly desired, especially due to the rising energy costs. Such ability to control illumination levels is very important for establishments that require a great deal of lighting, such as restaurants and offices.

[ooo4] Lighting levels that are higher than necessary not only result in a higher energy costs associated with the lighting, but also can increase air conditioning costs due to the excess heat provided by the lighting fixtures. Fluorescent light fixtures output less heat than incandescent light fixtures for equivalent illumination, and thus they are becoming more popular with offices or other commercial establishments. [ooo5] There currently exist various types of dimmer devices that can be used in order to control the amount of light output by fluorescent lights. One type utilizes a complex electronic ballast which first converts the applied AC line voltage to DC, then switches the applied tube voltage at high frequency. The resulting power-to-light output efficiency is hampered by this additional manipulation. This type requires an expensive fixture replacement and rewiring to the wall switch. Simplistic phase control devices will not provide satisfactory results when controlling a magnetic ballast fluorescent fixture.

[ooo6] Figure IA shows the connections of a conventional fluorescent dimmer device or controller 100, which is provided between a line and a load. The load is shown as a light fixture 110, which may be a fluorescent tube and associated ballast, for example. As shown in Figure IA, the conventional controller 100 needs access to both sides (line 102 and neutral 104) of an AC power input, in addition to the load. Since connectivity to the neutral line 104 is not always available at a light switch box, conventional fluorescent controllers may require expensive re-wiring to be installed. [ooo7] The problem with using such a conventional dimmer circuit for a fluorescent lighting fixture is that the conventional dimmer circuit cannot modulate reactive loads. Reactive loads react with the controller, thereby producing oscillations that then cause surges of voltage and current, which are both unpredictable and uncontrollable. With such control being applied to a fluorescent light fixture, the typical result is a non-harmonic type of flickering, which frequently takes the light from zero output to maximum output and to values in between. Such flickering is visually (and also audibly) discomforting, and may even be unhealthful to people who are near the flickering fluorescent light (for example, it may cause headaches due to having to view the undesirable light flickering).

[0008] As explained earlier, a controller such as the one shown in Figure IA can be used to control a fluorescent light without causing significant flickering, but such a controller requires fairly substantial installation costs, since they cannot be installed at a light switch box (where a neutral line is not typically provided), but rather have to be installed very close to the ballast (e.g., in the ceiling of a room, where a neutral line is provided). [ooo9] U.S. Patent No. 5,043,635 to Talbott et al. describes a two-line power control device for dimming fluorescent lights, which does not require to be coupled to a neutral line. Accordingly, the Talbott et al. device can in theory be installed at a light switch box. However, due to the analog structure and the various components described in the Talbott et al. device, such a device is very difficult to manufacture, and also such a device is very difficult to manufacture in a small size. Thus, it is not feasible to install such a device in a light switch box, given the bulkiness as well as the transformer configuration of the Talbott et al. device.

SUMMARY OF THE INVENTION

[0010] The present invention is directed to an apparatus and a method for controlling an amount of power supplied to a resistive, inductive or capacitive load by modulating a period of time that current flows through the load. toon] According to one aspect of the invention, there is provided an energy savings device for an inductive, resistive or capacitive load that is powered by an input AC voltage waveform. The device includes a setting unit for setting a desired power operating level for the load. The device also includes a processor configured to receive a signal from the setting unit indicative of the desired power operating level for the load, to determine a phase delay to be provided to an output AC voltage waveform that is to be provided to the load, and to output a control signal as a result thereof. The device further includes a synchronization circuit for synchronizing to a Safety Ground line, the synchronization circuit providing a synchronization signal that is used in construction of the control signal by the processor.

[ooi2] According to another aspect of the invention, there is provided an energy savings method for an inductive, resistive or capacitive load that is powered by an input AC voltage waveform. The method includes setting a desired power operating level for the load. The method further includes receiving a signal indicative of the desired power operating level for the load, and determining a phase delay to be provided to an output AC voltage waveform that is to be provided to the load, and to output a control signal as a result thereof. The method also includes synchronizing to a Safety Ground line, the synchronization circuit providing a synchronization signal that is used in construction of the control signal. [ooi3] According to yet another aspect of the invention, there is provided a computer program product for providing energy savings for an inductive, resistive or capacitive load that is powered by an input AC voltage waveform. The computer program product includes first computer code configured to set a desired power operating level for the load. The computer program product also includes second computer code configured to receive a setting signal output from the first computer code that is indicative of the desired power operating level for the load, the second computer code further configured to determine a phase delay to be provided to an output AC voltage waveform that is to be provided to the load, and to output a control signal as a result thereof. The computer program product further includes third computer code configured to synchronize to a Safety Ground line, to thereby provide a synchronization signal that is used in construction of the control signal by the second computer code

[ooi4] According to yet another aspect of the invention, there is provided an energy savings device for an inductive, resistive or capacitive load that is powered by an input AC voltage waveform. The energy savings device includes setting means for allowing a user to set a desired power operating level for the load. The energy savings device also includes processing means for receiving a signal from the setting unit indicative of the desired power operating level for the load, and for determining a phase delay to be provided to an output AC voltage waveform that is to be provided to the load, and to output a control signal as a result thereof. The energy savings device further includes synchronization means for synchronizing to a Safety Ground line, the synchronization means providing a synchronization signal that is used in construction of the control signal by the processing means.

BRIEF DESCRIPTION OF THE DRAWINGS

[ooi5] The foregoing advantages and features of the invention will become apparent upon reference to the following detailed description and the accompanying drawings, of which:

[ooi6] Figure IA shows a hookup of a conventional energy savings device that is provided between an input voltage line and a load; [0017] Figure IB shows a hookup of an energy savings device according to an embodiment of the invention that is provided between an input voltage line and a load;

[0018] Figure 2 shows an alternative hookup of an energy savings device according to an embodiment of the invention that provides neutral side control;

[ooi9] Figure 3 is a block diagram of an energy savings device according to a first embodiment of the invention;

[0020] Figure 4 is a schematic circuit diagram of an energy savings device according to the first embodiment of the invention; [oo2i] Figure 5 shows phase control waveforms according to the first embodiment of the invention;

[0022] Figure 6 is a software flow diagram of microprocessor firmware that operates according to the first embodiment of the invention;

[0023] Figure 7 is a block diagram of an energy savings device according to a second embodiment of the invention;

[0024] Figure 8 is a schematic circuit diagram of an energy savings device according to the second embodiment of the invention;

[0025] Figure 9 is a schematic circuit diagram of a master unit according to a seventh embodiment of the invention; [0026] Figure 10 is a schematic circuit diagram of a follower unit according to the seventh embodiment of the invention;

[0027] Figure 11 shows a synchronization circuit utilized in the universal control device according to an eighth embodiment of the invention;

[0028] Figure 12 shows a synchronization waveform as compared to the Hot AC line input voltage, according to the eighth embodiment of the invention; [0029] Figure 13 shows a circuit diagram of a universal control device according to the eighth embodiment of the invention being used to control a compact fluorescent lamp;

[0030] Figure 14 shows switching element trigger timings for the universal control device according to the eighth embodiment of the invention; and

[oo3i] Figure 15 shows a detailed schematic circuit diagram of the components making up the universal control device according to a preferred implementation of the eighth embodiment of the invention.

DETAILED DESCRIPTION OF THE INVENTION

[0032] Preferred embodiments of the invention will be described in detail below, with reference to the accompanying drawings.

[0033] The invention is directed to an apparatus and method for controlling power to a resistive, an inductive or a capacitive load, such as a fluorescent light fixture, a halogen light fixture, or a motor for a fan. In a preferred configuration, the energy controlling apparatus is configured to be installed in a light switch box typically located on an interior wall of a building, behind a wall switch plate. Since most light switches are mounted within a switch box that is easily accessible through the wall (e.g., behind a switch plate), the line to the switch is dropped from the fixture to the switch, and the other side of the line (e.g., neutral) is not conveniently present. The invention provides a true switch replacement and operates in series with an inductive or resistive load, in a two-wire configuration, plus safety ground wire. Figure IB shows a hookup of an energy savings device 150 according to an embodiment of the invention that is provided between the input AC line voltage 102 and a reactive load 110, whereby hookup to the neutral line 104 is not required by the energy savings device 150 in order to provide an energy control function for the load 110.

[0034] Additionally, referring now to Figure 2, some installations will wire the line 102 directly to the light fixture 110, leaving the load return 103 for fixture control. In this case, there is no line 102 connection in the switch box, again disallowing integration of a conventional fluorescent dimmer device. The UCD controller 150 is fully compatible with neutral 104 side control, in the manner as shown in Figure 2. In summary, the UCD controller according to the different embodiments of the invention is installed in series with the load, on either side of the load, without regard to wiring polarity, identically to a dry contact switch installation. [0035] With regards to fluorescent light fixtures, the energy savings device according to the invention regulates a voltage output to gaseous discharge lamps of the fluorescent light fixture from the secondary coils of a ballast element of the fluorescent light fixture. [0036] A universal control device (or UCD) according to a first embodiment of the invention will be described below in detail. A block diagram of the UCD according to the first embodiment is shown in Figure 3, and a schematic circuit diagram of the UCD according to the first embodiment is shown in Figure 4. [0037] The UCD according to the first embodiment includes a "push" On/Off switch and potentiometer unit 310 that is coupled to a line input (AC input voltage) 305, a solid state switch unit 320, a driver 330 for driving the solid state switch unit 320, a power supply 340, a microprocessor 350, and a line synchronization detector 360. The solid state switch unit 320 is provided between the line input 305 and the load 365. The switch and potentiometer unit 310 includes a "push" On/Off switch SWl and a potentiometer POT. The line synchronization detector 360 provides an interrupt signal to the microprocessor 350, which corresponds to "rising" zero crossing of a load current waveform, to be explained in more detail below.

[0038] The UCD is a two wire dimmer unit, and can be utilized to control standard magnetic fluorescent fixtures. The UCD may also be used to control other resistive, inductive or capacitive (e.g., standard electronic fluorescent fixtures) loads. The UCD functions similar to incandescent dimmers, but it also implements line synchronization functions and timing functions (not done by incandescent dimmers) to allow it to control fluorescent fixtures and/or other types of reactive or capacitive loads. In a preferred configuration, the UCD is wired in series with the fluorescent load without observance of wiring polarity, in either the hot or return side of the load, in a manner that is identical to a standard single pole wall switch. In fact, the UCD is configured so as to replace any existing wall switch to provide a dimming functionality. [0039] In a preferred implementation of the first embodiment, the UCD implements an 8-bit digital microprocessor 350 (of course, other types of microprocessors, such as 16-bit, 32-bit, etc., may be utilized instead of an 8-bit microprocessor, while remaining within the scope of the invention) with embedded firmware control algorithms for minimum parts count, and highly stable operation. The UCD according to the first embodiment is compatible with any configuration of magnetic ballast or electronic ballast fluorescent and/or incandescent loads. In a preferred construction, unit size, costs, producibility, performance and stability are optimized through the use of advanced digital and mass production techniques. Other embodiments to be described later include occupancy sensing, ambient light correction, and AC line modem for communication with a remote Energy Management System. All of the embodiments to be described herein are "in series", two wire devices (see Figure IB or Figure 2). [0040] Table 1 provides line specifications of the UCD according to a preferred implementation of the first embodiment of the invention. One of ordinary skill in the art will recognize that other line specification ranges may be handled by the

UCD according to the first embodiment, while remaining within the scope of the invention.

[0041] Table 1

[0042] The UCD according to the first embodiment provides AC line synchronization and timing firmware algorithms used to provide stable dimming control of an inductive and/or resistive and/or capacitive load without regard to applied line voltage, frequency, and without requiring a specific connection to the AC Line Return or Safety Ground. The UCD according to the first embodiment implements phase control of the load, and also strategically controls the switching element turn-on timing for stable (non-flickering) control of inductive or resistive loads. The UCD according to the first embodiment synchronizes on the load current zero crossing, which causes a turning off of the series switching elements making up the solid state switch unit 320.

[0043] Highly inductive or resistive loads, such as magnetic fluorescent ballasts, cause a significant phase shift (delay) of the load current waveform relative to the applied voltage waveform, greatly complicating stable synchronization. This phase shift varies depending on the specific installation (number of fixtures and specific ballast specifications) as well as the selected dimming level. As the dimming level is varied, or fluorescent tube temperature changes, the current zero crossing synchronization signal to the microprocessor will move significantly in real time, causing a shift in phase timing for the next cycle. Unless a suitable phase timing algorithm is implemented, the light fixture will flicker in an oscillatory way, resulting in unstable (highly unsatisfactory) dimming. The inventors of this application realized that standard incandescent dimmers will not reliably function with fluorescent or other types of reactive loads due to their simplistic line synchronization methods. The timing correction algorithms utilized in the present invention are an important aspect of the UCD design according to the first embodiment (as well as to the other embodiments), and are described in detail below. Also, the UCD according to the present invention also performs well as a dimmer control with little or no flickering, for an electronic fluorescent ballast, which is a capacitive load.

[0044] Figure 5 shows the applied line voltage waveform, the dimmed fluorescent load current waveform, and the microprocessor synchronization waveform as implemented by the UCD according to the first embodiment. Also shown in Figure 5 are seven (7) time points in a single cycle of the applied line voltage waveform (60 Hz or 16.67 msec time period for one cycle), each of which is discussed in detail below. The highly inductive nature of a fluorescent magnetic ballast causes the load current to lag the applied line voltage, as seen in the comparison of the AC line voltage waveform 510 with the load current waveform 520. The amount of lag depends on the circuit inductance, specific ballast design factors, tube striking voltage which is affected by tube temperature, and the amount of dimming phase delay being applied by the UCD according to the first embodiment. A point by point discussion of the seven labeled time points in Figure 5 follows, with reference to the circuit elements shown in Figure 4. [0045] Time point 1 corresponds to the rising zero crossing of the applied line voltage waveform 510. [0046] Time point 2 corresponds to the turn off point of Silicon Controlled Rectifier (SCR) Q2 from the previous dim cycle. An SCR turns off when the applied current through it reaches zero. Once the SCR turns off, the voltage across the SCR rises sharply.

[0047] At time point 3, the turning off of SCR Q2 causes the synchronization signal on pin 5 of the microprocessor U2 to go low, which interrupts the microprocessor U2. In the preferred implementation of the first embodiment, microprocessor firmware is initialized to only respond to the falling edge of the interrupt, and is used to derive all phase control timing for an entire line cycle. As the UCD dimmer potentiometer R7 is rotated clockwise, the period of phase delay time between time point 3 and time point 4 of Figure 5 is increased, causing the fluorescent light fixture being controlled by the UCD to dim. Conversely, counterclockwise rotation of the UCD dimmer potentiometer R7 decreases this phase delay time, thereby causing the fluorescent light fixture light output amount to intensify. [0048] The inventors have found through experimentation that a typical fluorescent tube with magnetic ballast goes off (no light output by it) at approximately 120 degrees (about 5.5 mseconds) of phase delay. This is due to insufficient tube ionization caused by insufficient tube heater output. Without adequate tube ionization, the tube strike voltage exceeds that available from the AC line. The inventors have also found that they were not able to visibly discern a change in light output until the phase delay reached about 15 degrees (about 0.7 mseconds) of phase delay. The half-intensity point was about 90 degrees of phase delay (about 4.17 mseconds).

[0049] Microprocessor control of the phase delay controls the dim level of the fluorescent fixture (the load). In response to the falling edge of the synchronization interrupt, the microprocessor U2 resets a free-running internal hardware timer (not shown in the figures) to zero, then waits for the timer to reach the phase delay value corresponding to the current position of the UCD dimmer potentiometer R7. In a preferred implementation, the UCD dimmer potentiometer R7 is coupled to a rotatable dial that is disposed on a wall of a building, whereby, when a user rotates the dial, the resistance of potentiometer R7 changes accordingly. The change in the resistance of potentiometer R7 is discerned by the microprocessor U2, which then computes a different phase delay value for a next AC voltage waveform cycle based on the new dimmer setting. [0050] After the calculated phase delay time corresponding to time point 4 is reached, the microprocessor U2 triggers the SCR Ql on by bringing pin 2 of the microprocessor U2 low for a short period of time. In the preferred construction, an opto-isolated triac Ul is used to trigger the SCR on while isolating the microprocessor U2 from possible damaging transients. Once the SCR Ql is triggered on and current begins to flow, SCR Ql will latch itself on until current reaches zero during the next half cycle. Current flow through the load continues whenever the SCR Ql or the SCR Q2 is triggered on. When the SCR triggers on, the synchronization signal 530 goes high again. The rising edge of the synchronization signal 530 is ignored by the microprocessor U2, which only reacts to a falling edge of the synchronization signal 530 (due to microprocessor firmware that allows interrupts only on the falling edge of a signal provided to its interrupt port).

[oo5i] Time point 5 corresponds to the next zero crossing of the load current waveform 520. At this point, the SCR Ql turns off. Unlike the occurrence at time point 2, no synchronization signal occurs at time point 5. This is because the microprocessor 5V supply voltage (input line voltage) 340 is negative (it is a floating supply), and the open fluorescent circuit (that is, the load) is roughly ground. The synchronization signal 530 actually rises slightly (few tenths of a volt) after time point 5, because the "grounded" fluorescent circuit is actually higher in voltage than the microprocessor negative 5 V power supply 340. Microprocessor firmware is provided such that no microprocessor interrupt is generated from this slight perturbation of the synchronization signal 530 (and also since it does not correspond to a voltage drop but rather a voltage rise). too52] Phase control for the latter half-cycle of the AC line voltage waveform 510 is derived from the previous earlier half-cycle interrupt. The microprocessor U2 measures the applied line frequency and computes the number of internal free- running hardware timer counts that it has to wait for before triggering the SCR on for this latter half-cycle. The timer counts for a time period corresponding between the time between time point 5 and time point 6. [0053] At time point 6, the SCR Q2 is triggered on. At time point 6, the voltage of the synchronization signal 530 drops slightly (a few tenths of a volt). No microprocessor interrupt is generated here either, due to the microprocessor firmware being configured to not cause an interrupt for such a small voltage drop. Again, the SCR Q2 remains on during the negative half cycle, until the circuit current reaches zero.

[oo54] At time point 7, the rising load current waveform 520 again reaches zero. Again the synchronization signal 530 goes to zero, which causes a microprocessor interrupt (since it is a falling edge of the synchronization signal 530). This also causes a resynchronization of an internal free-running timer of the microprocessor U2, and results in another phase delay cycle similar to the one that was described above with respect to the time point 2 and time point 3. [0055] The UCD hardware design according to a preferred configuration of the first embodiment includes the components illustrated in the Figure 4 schematic diagram. A brief description of each hardware component, and its applied function, is provided below, loose] The microprocessor U2 (which corresponds to microprocessor 350 of Figure 3) provides the control functions and algorithms for the UCD according to the first embodiment based on an internally stored firmware program. By way of example and not by way of limitation, in a preferred implementation, a MICROCHIP™ 12C672 eight bit microprocessor incorporates 2 kilobytes programmable read only memory (PROM) for program storage, 128 bytes random access memory (RAM), an eight bit timer, 4 channel 8 bit Analog to Digital (AfD) converter, 4 MHz oscillator, and reset circuit in a very space efficient 8 pin package. More details on this microprocessor can be found at the Internet web site www.microchip.com. Of course, one of ordinary skill in the art will understand that other types and sizes of microprocessors may be utilized for the microprocessor to used in the first embodiment, while remaining within the scope of the invention.

[0057] Since the functionality of the microprocessor U2 exists internally, in a preferred implementation, six I/O pins may be allocated to either digital inputs and outputs or analog inputs. Two pins are reserved for +5 volt power and ground. By way of example and not by way of limitation, an Analog to Digital input impedance is approximately 1OK ohms.

[0058] By way of example and not by way of limitation, the "push" on/off potentiometer switch SWl is rated for the 6.3 ampere maximum dimming capacity. When turned off, the dimmer/load is entirely open circuited, resulting in no current flow to the load. Rotating potentiometer R7 and switch SWl are preferably integrated into a single unit. Pushing the adjustment shaft of potentiometer R7 will cycle switch SWl on and off. Potentiometer R7 is wired as an adjustable voltage divider, whereby rotating the shaft of potentiometer R7 adjusts the voltage at pin 7 of microprocessor U2. The microprocessor U2 reads the voltage at its pin 7 once every AC line cycle, and uses this voltage to derive the amount of phase delay (dim level) for the load. Resistor R8 is wired between the potentiometer wiper and ground, and is used to provide a more linear relationship between the potentiometer position and resulting dim level. By way of example and not by way of limitation, resistor R8 has a resistance of 4.7 kohms. [0059] In the preferred implementation of the UCD according to the first embodiment, two SCRs Ql, Q2 are connected back to back to provide an active switching element for the UCD, and correspond to the solid state switch 320 of Figure 3. The inventors found that TRIAC devices do not trigger as accurately as back-to-back SCRs when switching a highly inductive resistive load. Consistent and accurate switching element turn-OFF at the current zero crossing is very important to line synchronization. The use of a TRIAC as the active element may result in occasional flickering, which may be due to an unstable holding current level. As a result, the inventors found that an active element that includes back-to- back SCRs functions much better than one having a TRIAC in the energy savings device according to the invention, whereby using two SCRs provides an increase in switching current capability and better heat distribution to a heat sink. [0060] By way of example and not by way of limitation, the SCRs utilized in a preferred implementation of the first embodiment are 600V₅ 15 ampere devices. The SCRs Ql, Q2 are designed to run very cool at maximum specified loads. The choice of which type of SCRs to use in the first embodiment may also be made based on a low holding current parameter for the SCRs. When a signal of either polarity triggers the opto-isolated triac Ul, positive pulses from pin 4 and from pin 6 of the opto-isolated triac Ul are transmitted to gates (G) of the SCRs Ql₅ Q2, respectively. Opto-isolated triac Ul of Figure 4 corresponds to solid state driver unit 330 shown in Figure 3.

[oo6i] SCRs conduct current in one direction (from anode to cathode), with back- to-back SCRs having the capability to conduct in both directions. SCRs are latching devices, meaning that once they are trigger on, they will continue conducting until the anode-to-cathode current through them reaches zero (or reverses direction). An SCR is triggered on by pulling current out of its Gate pin, or bringing the Gate voltage a few volts lower than its anode pin. The holding current specification for an SCR specifies the minimum SCR current necessary for the SCR to latch on, and to remain latched on. A holding current on the order to 20 milliamperes is needed for proper operation of a typical SCR. Once the SCR current drops below the specified holding current, it will turn off until retriggered again. Only the SCR with its anode voltage positive relative to its cathode voltage is capable of being triggered on. This means that SCR Ql controls the load during the positive half of the AC voltage waveform cycle, and SCR Q2 controls the load during the negative half of the AC voltage waveform cycle. [0062] As shown best in Figure 4, the opto-isolated triac Ul is used to trigger the SCRs Ql, Q2. The microprocessor U2 triggers opto-isolated triac's Ul internal triac, and subsequently one or the other SCR Ql, Q2, by illuminating the opto- isolated triac's Ul internal light emitting diode (LED). LED illumination occurs when the microprocessor U2 pulls its output pin 2 low, resulting in LED forward current. The opto-isolated triac Ul is capable of conducting current in either direction, depending on the relative voltages of pins 4 and pin 6 of the opto- isolated triac Ul. For example, if pin 6 is higher than pin 4 of the opto-isolated triac Ul, current will flow from pin 6 to pin 4. Connecting the opto-isolated triac Ul between the gates of the two SCRs Ql, Q2 provides a convenient method of triggering back-to-back SCRs.

[0063] Current flows into pin 6 of the opto-isolated triac Ul and out pin 4 in response to the positive half of the AC sine wave voltage waveform 510 (see Figure 5) and vice versa in response to the negative half of the AC sine wave voltage waveform 510. Pulling current out of the associated SCR gate turns the device on. The internal structure of the SCR allows current to flow into the gate of the opposite device without triggering the device. Therefore, SCR Ql will remain latched through the positive half of the sine wave current, whereupon at approximately zero crossing, the latching current will be insufficient and SCR Ql will switch off. Similarly, the gate of SCR Q2 will source current into pin 4 of the triac Ul and out pin 6 of the triac Ul during negative half of the AC cycle, and remains latched again until approximately zero crossing. This switching sequence repeats for each cycle of the AC sine wave voltage waveform 510, providing full power of sine wave current to the (fluorescent) load. Accurate and stable triggering of the SCRs Ql and Q2 are very important to the suppression of flickering. [0064] Back-to-back SCRs are used to form an active element of an energy savings device according to a preferred implementation of the first embodiment since they were found by the inventors to be somewhat more stable in their turn OFF characteristics than a TRIAC. In order for an SCR to latch on, the anode/cathode current must exceed the latching current requirement. Once it is latched on, an SCR will remain on until it is turned off when anode/cathode current drops below holding current requirement. With such features, SCRs are ideal devices to be utilized for the active element that corresponds to the solid state switch 320 (see Figure 3) of the UCD according to the first embodiment. One of ordinary skill in the art will recognize that other types of solid state switches may be utilized, as well as switch drivers, beyond the ones described herein, while remaining within the scope of the invention. cooes] In the preferred implementation of the first embodiment, the opto-isolated triac Ul is utilized to provide driving signals to the SCRs Ql, Q2. By way of example and not by way of limitation, the opto-isolated triac Ul may be a MOC3022 opto-isolated triac, which drives the Ql and Q2 gates and provides line transient protection to the microprocessor U2. A LED drive current of approximately 5 milliamps (via resistor R6, which is a 620 ohm resistor in the preferred implementation) is sufficient to reliably trigger the opto-isolated triac Ul . The GP5 pin of microprocessor U2, which corresponds to pin 2 of the microprocessor U2, is configured for output and is capable of sinking up to 20 milliamps. [0066] Referring to Figure 5, the opto-isolated triac Ul outputs a drive signal starting at time point 6, whereby the drive signal is turned off well before the load current zero crossing at time point 7. Also, the opto-isolated triac Ul outputs a drive signal starting at time point 4, whereby the drive signal is turned off well before the load current zero crossing at time point 5. [0067] Referring to Figure 4, resistor R2 is a current limiting resistor, and is provided so as to limit the series current of the opto-isolated triac Ul to be less than one ampere under all circumstances. For 277 VAC installations, the value of resistor R2 should preferably be increased to 470 ohms due to the increase in the AC waveform voltage level. [0068] In a preferred implementation of the first embodiment, the SCR trigger signal output by the optoisolated triac Ul stays on for approximately 1.2 milliseconds. The actual SCR trigger signal on time is not critical, since an SCR triggers on within a few microseconds of receiving a trigger signal to its gate. In a preferred implementation, and as explained above, the SCR trigger signal turns off before the next zero crossing of the load current waveform, in order to enforce some SCR off time (e.g., 0.25 milliseconds). This off time is provided in order to recharge the 5 volt power supply 340 (see Figure 3) for the next cycle. [0069] Resistor Rl, capacitors Cl, C2, diodes Dl and D2, and the 5 volt power supply of Figure 4 are all utilized for a power supply control for the UCD according to the first embodiment, and together form the power supply unit 340 shown in Figure 3. In a preferred implementation, the 5 Volt power supply 340 provides up to 20 millamps of power to the microprocessor U2, opto-isolated triac Ul , and the potentiometer R7 at all times in which the UCD is powered. The 5 Volt power supply 340 floats with the AC line input. Voltage is derived by the widely varying voltage across SCRs Ql and Q2. Power is available to the circuit only when SCRs Ql and Q2 are switched OFF. When SCRs Ql and Q2 are turned on, the 5 Volt supply 340 is maintained by capacitor Cl and is stabilized by zener diode Dl . Silicon Diode D2 provides a discharge path for capacitor Cl . Resistor Rl and capacitor C2 provide an AC coupled voltage drop to limit silicon diode Dl and zener diode D2 current and dissipation. By way of example and not by way of limitation, the microprocessor U2 remains entirely functional with any supply voltage over 3.3 Volts at a current of 3 milliamps. In a preferred implementation of the first embodiment, supply regulation is not critical as long as the supply voltage maintains the 3.3V minimum.

[0070] Resistors R3, R2, R4, R5, and diode D3 of Figure 4 are elements making up the Line sync unit 360 shown in Figure 3. The falling half of the AC line output (when SCRs Ql and Q2 turn off) is used for line synchronization. SCRs Ql and Q2 turn off at the line current zero crossing. Zener diode D3 protects the microprocessor interrupt input (port 5 of the microprocessor U2) against unforeseen line and switching transient spikes. Resistor R5 limits current input to the microprocessor U2 and allows the internal microprocessor protection or clamp diodes to function while preventing any possible burnout. Resistors R2, R3 and R4 also provide a current limiting and line synchronization function for the UCD. [0071] The inventors have realized that stable AC line synchronization is very important to non-flickering operation when controlling inductive and/or resistive loads (especially conventional Magnetic Ballast Fluorescent Fixtures), and even for controlling capacitive loads (such as Electronic Ballast Fluorescent Fixtures). These synchronization methods are implemented in the firmware of the microprocessor U2 according to the first embodiment, and are applicable to the other embodiments as well.

[0072] The microprocessor firmware provides a Line Sync Edge Detection function. In detail, the microprocessor U2 is interrupted on the falling edge of Line Syncronization signal 530 (see Figure 5) which occurs once every AC cycle as the switching element turns off at the current zero crossing. SCRs have a characteristic in that they latch themselves on until the current through them reaches zero. The point where they turn off is used as the line synchronization. An internal timer of microprocessor U2 is initialized at this interrupt, and timing parameters for the next entire AC cycle calculated in firmware. Using a single current zero crossing per AC cycle cancels any non-uniformity of the positive and negative halves of the current waveform, as well as eliminates interrupt input threshold hysteresis effects.

[0073] The firmware of microprocessor U2 also provides an AC Line Period Determination function. In detail, at initial power up, the microprocessor performs a timing analysis of the AC line with the load switched off so that specific timer counts for each half phase may be calculated. Leaving the load off during this period provides a very accurate measurement of the AC line voltage, without inductive load phase shift influence. At the first interrupt after initial power up, the microprocessor timer is initialized to zero. At the next interrupt the timer value is stored, representing the number of timer counts for a full AC cycle. Subsequent phase timing parameters are derived from this number. Intra-interrupt timing functions are driven by waiting for specific timer counts. [0074] The microprocessor firmware also performs a Phase Timing Calculation function. In detail, once the line period has been determined, the firmware of microprocessor U2 performs phase timing calculations. Since synchronization is performed only once per AC cycle, a determination of the cycle half time is made by dividing the period by two (shift right one time). Next, a calculation of when the cycle is completed (cyclendtime) in anticipation of the next interrupt is made. [0075] The firmware of microprocessor U2 further performs a Dead Time Implementation function. In detail, circuit power is only available when the series switching elements (SCRs) are turned off, therefore microprocessor firmware guarantees a minimum off time (deadtime) for each AC line half cycle to restore the 5 volt supply.

[0076] The firmware of microprocessor U2 also performs a Fixture Warmup function. In detail, fluorescent tubes should be fully warmed up before they can be reliably dimmed. This feature may not be desirable for other types of inductive or resistive loads, and may be easily deleted from the control device, without departing from the scope of the invention. To address this requirement, the fixture is set to full intensity for a first time period after initial power up. By way of example and not by way of limitation, the first time period is set to 12 seconds. Upon completion of the 12 second period, the intensity is returned to the dim level corresponding to the position of potentiometer R7 (see Figure 4). [0077] The firmware of microprocessor U2 further provides a Sync Window

Implementation function. In detail, in order to reject spurious line transients which could possibly upset dimmer timing, a sync window algorithm is utilized in the first embodiment. At the end of each full AC cycle, the microprocessor U2 waits until cyclendtime which occurs a few timer counts before the next line interrupt, before re-enabling interrupts. If a spurious interrupt occurred between the last sync edge and cyclendtime, it is effectively ignored. [ooze] The firmware of microprocessor U2 also provides a Slow Phase Timing (Dim Level) Changes function. In detail, when using a current zero crossing sync with an inductive magnetic ballast, any phase timing (dim level) change causes a slight synchronization variance which could cause instability (flickering) if not greatly damped out. To greatly lessen this possibility, phase timing changes are limited to one timer count per AC cycle, thereby minimizing this effect.

[0079] The firmware of microprocessor U2 further provides a function for pulsing the SCRs ON at the correct time. In detail, the SCRs Ql, Q2 are pulsed on, instead of just turned on and left on at the proper time, to reduce the drain on the 5 Volt power supply 340 (see Figure 3). [0080] More details of the microprocessor firmware implementation according to a preferred implementation of the first embodiment is provided in detail below. In the preferred implementation, the firmware of microprocessor U2 is written using a Microchip assembler language specific to the 12C672 eight bit microprocessor. Of course, based on the type of microprocessor utilized in the first embodiment, the choice of software language used to write the microprocessor firmware will be utilized accordingly.

[0081] A detailed flow chart of the preferred implementation of microprocessor firmware to be utilized by a microprocessor U2 according to the first embodiment of the UCD is illustrated in Figure 6. Major flow chart function descriptions are provided below.

[0082] For UCD implementation, a Reset occurs only during initial power up. At this time, microprocessor memory and register contents are random, and are thereby initialized before they can be used. In the preferred implementation of the first embodiment, the microprocessor U2 has an internal reset circuit which recognizes when power is initially applied. Upon Reset, the microprocessor U2 begins execution at address 0000, which is where the initialization firmware starts. Once this initialization executes, it is not re-executed unless another power up sequence occurs.

[0083] Two interrupts are enabled for the UCD according to the first embodiment. First, the external synchronization falling edge interrupt, from which all phase delay calculations are derived, is enabled. Second, the internal hardware free- running timer overflow interrupt is enabled. In the preferred implementation of the first embodiment, the timer is an 8 bit timer which is incremented once every 64 microseconds. The timer overflows every 16.384 milliseconds (256 counts), which is slightly less than a full 16.667 millisecond line cycle. During an interrupt, the microprocessor U2 stops executing where it is, saves it's state (e.g., processor status word and program counter), and executes interrupt code. Initial line parameter calculations, hardware timer maintenance, and Analog to Digital Converter (ADC) maintenance occurs during the interrupt firmware. [0084] Referring to Figure 6, "Main" is the start of the primary UCD software program run by the microprocessor U2. It is entered after initial power up initialization and once per complete line cycle. "Main" keeps track of the current line half cycle, and performs all phase timing calculations based on the free- running hardware timer. Phase timing is implemented by waiting for the appropriate free-running timer count to occur, then calling the TrigScr subroutine which implements the SCR trigger timing. Specific free-running timer values to wait for are calculated based on the following factors: [0085] a) Dimpot position: As indicated by the converted ADC value. Rotating the dimpot potentiometer clockwise will reduce phase delay, and increase florescent intensity.

[0086] b) FullOnMode: During the first 12 seconds after initial power up, the UCD is in FullOnMode. During this time, the florescent load is forced into full intensity to warm the tubes. During FullOnMode, phase delay is fixed at the constant value fulltime. When not in FullOnMode, phase delay is calculated based on dimpot position, and results of the softdim calculation. The softdim calculation prevents large cycle to cycle phase delays from occurring. This provides a stabilizing effect on florescent intensity.

[0087] c) Cycle Half: After completion of the first half of the line cycle, firmware waits for the pre-calculated half period free-running hardware timer value, resets the timer, and jumps back to Main. This causes the second half cycle phase delay timing to be identical to the first half cycle. At the end of the second half cycle, firmware will wait for the free-running hardware timer to reach the pre- calculated cyclendtime, then re-enable interrupts in anticipation of the next full line cycle.

[0088] After the appropriate phase delay has be determined, a call to TrigScr is executed whereby the SCRs Ql, Q2 are turned on at the appropriate times. [0089] The TrigSCR sub-routine toggles the SCRs Ql, Q2 on and off for a period of time to minimize drain on the 5 V power supply. Once the SCR current is greater than the SCR specified holding current, it will latch on for the duration of the half cycle, until the current reaches zero again. Relative free-running hardware timer values are used to accomplish this pulse ON, pulse OFF, and pulse duration timing.

[0090] The following are descriptions of each section of the dimmer firmware utilized by the microprocessor U2 according to a preferred implementation of the first embodiment, whereby each section is identified by line number, then label and references to the flow chart of Figure 6. Of course, other firmware may be utilized as would be recognized by one of ordinary skill in the art, while remaining with the scope of the invention. [0091] Line 1 : Defines the microprocessor as the target for the assembler

[0092] Line 2: This include file defines the microprocessor register names and memory mapped register addresses.

[0093] Line 5: A list of defined memory mapped addresses follows: dimpot: Storage of the dim potentiometer analog value timerstat: Mode Flags specific to dimming mode tmrovflcntr: Used as an overflow counter to the internal 8 bit counter

TMRO intovflcntr; LSB of counter used for 12 sec full ON fullintcntr: MSB of counter used for 12 sec full ON timereg: Temp Storage of TMRO Count periodmsb: Measured MSB of Full wave TMRO Count periodlsb: Measured LSB of Full wave TMRO Count halftime: Calculated TMRO Count for Half Wave trigtime: Calculated TMRO Count to Trigger SCR SCRofftime: Temp Storage where time to turn off SCR is stored each cycle

SCRlstime: Temp Storage for Last SCR time... subsequent SCR ON/OFF functions key off of this stored TMRO value cycendtime: Re-Enable Edge Interrupt time softlast: Temp Storage of last dim time count is stored. Used for

Soft Dim [0094] Line 23 ;GPIO Bit Defs potanal 12C672 GPIO Pin Allocated to Potentiometer Analog Input gpl 12C672 GPIO Pin Not Used acint 12C672 GPIO Pin Allocated for AC Interrupt Input gp3 12C672 GPIO Pin Not Used gp4 12C672 GPIO Pin Not Used

SCRdrv 12C672 GPIO Pin GPIO SCR Drive Output [0095] Line 31 ;TimerStat Bit Defs firstedg Flag: First Interrupt Edge Occured secedge Flag: Second Interrupt Edge Occured fullonmode Flag: Full on mode newedge Flag: New Edge Flag cycsechalf Flag: Second Half of Period oddedge Not Used in this Version [0096] Line 39 ; Value Defs intovflow = d'3' ;FullOnMode Int Overflows ~4Secs per inc dimofst = h'4' ;ADC Offset, Higher Numbers go Dimmer maxofst = h'7f ;Maxdim Offset maxdima = h'fe¹ ;Maxdim Level maxdimlvl = h'dO' ;Maxdim intwindow = d'3' interrupt Window

SCRpulsetime = h'37' ;Time SCR is Pulsed ON and Off deadtime = d'8' ;Dead time past zero crossing fulltime = d'8¹ ;Full On time past zero crossing [0097] Line 54 rstvec The microprocessor starts execution at address 0 after Reset, Interrupts are disabled, then memory initialized [0098] Line 58 intvec The microprocessor interrupt vector for enabled interrupts is at address 4 [0099] Line 59 intsvc TMRO is cleared at each falling edge of the AC interrupt. After a Reset, a wait for the zeroth edge is executed. Upon occurrence of the zeroth edge, TMRO overflow interrupt is enabled so that the AC edge to edge period can be calculated. Upon occurrence of the first edge interrupt, AC parameters are calculated and used in subsequent phase calculations. [00100] Line 61 Jump table based on edge occurrences [00101] Line 65 notfirst Zeroth edge interrupt has occurred, enable TMRO overflow Interrupts [ooio2] Line 72 firsthap First interrupt has happened, count number of TMRO overflows, enable Next TMRO overflow interrupt [ooio3] Line 78 notmrint If it's a second edge interrupt, then disable subsequent TMRO overflow Interrupts, and then calculate AC timing parameters

[ooio4] Line 81 caltime AC parameters such as period, halftime, and cyclendtime, are calculated once. Flag secedge is then set, and further edge interrupts enabled. From now on, each edge interrupt constitutes an AC line synchronization signal used for phase control of the SCRs

[ooio5] Line 100 sechap Once the second edge interrupt has occurred, then 12 seconds of full on is executed to fully warm the tube heaters. Fullintcntr, and intovflcntr form a 16 bit counter which count 16.667 mS edge interrupts. A total of 768 edge interrupts provides a net 12.8 seconds of fluorescent tube full on time.

[00106] Line 112 fulldun Upon conclusion of the full on mode, the fullonmode flag is cleared in timerstat. [00107] Line 113 notfull Each edge interrupt, the A/D converter is checked for conversion complete. If it has completed the dimpot value is inverted by exclusive Oring the input value and stored in the memory location dimpot. [00108] Line 121 nocvrt A/D conversion has completed, another conversion is started. The newedge flag is set and the cycsechalf flag cleared, indicating to the main program code that an interrupt had occurred, and that it is now the first half of the AC cycle.

[ooio9] Line 123 glitint TMRO is cleared, Edge interrupts are re- enabled, and a return from interrupt executed

[00110] Line 129 initmem Microprocessor hardware registers are initialized, program defined registers are cleared, and finally edge interrupts are enabled.

[00111] Line 173 main Main part of the program. Wait for second edge interrupt. At this time, all AC line parameters have been calculated, and normal phase control can commence.

[ooii2] Line 175 mainl Wait for each new edge. Newedge is a handshake flag with intsvc which is used to wait for a new edge at the completion of each AC cycle. [ooii3] Line 178 main2 Entered at the start of each AC cycle.

Potentiometer scaling to actual TMRO counts are performed once per AC cycle. Edge Interrupts are disabled, dimpot contains the commanded dim value. The memory location softlast is used to calculate the desired dim value time. [ooii4] Line 189 sechalf This is the entry point for the second half of the AC cycle. IfNOT in Fullonmode, then go to dimtrig. Else, it is fullonmode at sechala. [00115] Line 191 secala A wait until TMRO = deadtime is executed.

Deadtime defines the earliest time (in TMRO counts) the SCR may be triggered ON after an AC line voltage zero crossing. A call to trigSCR turns the SCR on for a period of time. After returning, the first cycle half is complete.

[00116] Line 198 dimtrig Fullonmode has completed, enforce minimum deadtime limit, by waiting for TMRO to reach deadtime value.

[ooii7] Line 202 dimwait Past deadtime, now wait for the calculated TMRO value corresponding to the calculated phase delay for the indicated dim level. The memory location trigtime is incremented or decremented once each time, effectively "chasing" the desired dim level stored in softlast.

[00118] Line 217 hafcycl Halfcycle parameters are checked. If already in the second half, a wait for next edge interrupt (jump to rstcycle) is executed. If Not already in second half, a wait until the previously calculated Halftime TMRO value is executed. Once past halftime, TMRO is cleared, and the cycsechalf flag is set. Then a jump to sechalf occurs, duplicating timing parameters for the second half of the AC cycle.

[ooii9] Line 229 rstcycle Once timing for the second half of the

AC cycle has been executed, a wait until cyclendtime is executed before edge interrupts are Re-enabled. This provides a window which rejects AC line transients which occur outside of the window. Upon passage of the window, Interrupts are re-enabled, and a jump to mainl is executed, causing a Wait for the next edge interrupt. [00120] Line 240 trigSCR TrigSCR is a routine that is called when it's time to turn on the SCR. When called, the SCR is triggered on (SCRdrv is brought low), then the SCRofftime is calculated based on addition of the constant SCRpulsetime, and the current TMRO value. A wait until SCRofftime is executed, whereupon the SCR is turned off (SCRdrv is brought high). If cycendtime occurs during the time trigSCR executes, drive to the SCR is deasserted, and a return to the calling code is executed.

[00121] Line 265 end End of the program. [ooi22] Figure 7 shows a block diagram of an energy savings device UCD-2 according to a second embodiment, and Figure 8 shows a schematic circuit diagram of the energy savings device UCD-2 according to the second embodiment. The energy savings device UCD-2 according to the second embodiment provides all of the functions of the first embodiment, along with extra functions. The UCD- 2 includes an occupancy sensor, an ambient light sensor, and an AC line modem for remote communications to a central energy management system, for example. The UCD-2 provides a more robust energy savings function than the UCD according to the first embodiment.

[00123] As shown in Figure 7, an ambient light sensor unit 710 of the second embodiment provides the capability to adjust the dimming level for constant level illumination during day/night ambient illumination variances. Referring also to Figure 8, the ambient light sensor unit 710 includes a photo-resistor Rl 9 with amplifier 720, which provides a stable indication of the total ambient illumination via a signal AMBLITE provided to port 1 of the microprocessor U2. The microprocessor U2 adjusts the dimming level to maintain this total ambient illumination level. For example, during a cloudy day, if the clouds break during the afternoon and thus the light through windows of an office increases, this results in an increase in the illumination level picked up by the ambient light sensor unit 710. Accordingly, the microprocessor U2 will adjust the load current waveform to provide a slightly dimmer signal than what was previously provided (during the cloudy period), so as to maintain a stable ambient illumination for the office. tooi24] Referring to Figure 7 , the occupancy sensor unit 730 of the second embodiment provides the capability to sense movement within an illumination area. The occupancy sensor unit 730 is configured to provide a signal indicative of no movement to the microprocessor U2 if no movement is sensed after an extended interval of time (e.g., 15 minutes or more). Upon receipt of the "no movement" signal from the occupancy sensor unit 730, the microprocessor U2 turns the light fixture off, in order to save energy. Similarly, illumination to a preset level is restored if movement occurs, such as when a person walks into a room. Referring to Figure 8, the occupancy sensor unit 730 according to a preferred implementation includes a passive infrared sensor 750 with a multifaceted (Fresnel) lens 740 in front of a pyroelectric transducer. For example, a Murata IRA-E71 OSTO may be utilized as the motion detector for the occupancy sensor unit 730. The lens 740 focuses infrared energy from a multitude of narrow, discrete beams or cones. As a warm body moves across the field of view of the detector, the transducer output has peaks and valleys which are amplified, thereby providing an indication that movement is occurring. This results in a signal MOTDET that is indicative of movement being provided to the microprocessor U2.

[ooi25] Referring to Figure 7, the AC line modem 760 of the second embodiment enables bi-directional communications with an energy management unit, such as with a centralized energy management system (EMS). In one implementation shown in Figure 8, the AC line modem is implemented as a line modem TDA5051 component. The EMS has the capability to remotely control some or all dimming functions and modes including turn off illumination (via signal PWRDWN provided to microprocessor U2), set dimming level, and verity occupancy sensor status (possible burglar alarm function). The EMS is preferably a standard personal computer with external AC line modem connected to a serial port. Software running under an operating system, such as the Windows™ operating system, maintains the status of all units within a local area. The AC line modem 760 functions by modulating a 200 KHz signal onto the AC power line via a filter network 770 that includes an inductor Ll and a capacitor C4 (see Figure 8), in one possible implementation of the second embodiment. The EMS can communicate with a wide area of dimming units that are on a common AC line step down transformer, for example. Each dimming unit carries a unique address to facilitate a multi-drop communications network via the power lines. [00126] In a third embodiment, unlike the "loaded" second embodiment, only the ambient light sensor unit of the second embodiment is provided along with the features of the first embodiment. [ooi27] In a fourth embodiment, only the occupancy sensor unit of the second embodiment is provided along with the features of the first embodiment. [ooi28] In a fifth embodiment, only the AC line modem of the second embodiment is provided along with the features of the first embodiment. In another possible implementation, both the occupancy sensor unit and the AC line modem (but not the ambient light sensor) of the second embodiment are utilized along with the features of the first embodiment. In yet another possible implementation, both the AC line modem and the ambient light sensor (but not the occupancy sensor unit) are utilized along with the features of the first embodiment. In still yet another possible implementation, both the occupancy sensor unit and the ambient light sensor (but not the AC line modem) are utilized along with the features of the first embodiment. [00129] A sixth embodiment of the invention includes all of the features described above with respect to the second embodiment, as well as a remote control function. The remote control function allows a user to set a light level by a remote control unit, without having to go to a switch box on a wall. By pointing the remote control unit in a direction of the switch box, and by enabling a button on the remote control unit, a signal is picked up by an element (e.g., infrared sensor, IR sensor) on the switch box, similar to a television remote control unit, whereby a room light level is either increased or decreased depending on the user's selection on the remote control unit. The remote control function can also be used with any of the other embodiments described above.

[00130] A seventh embodiment of the invention is described herein with respect to Figures 9 and 10. The seventh embodiment is directed to a master/follower control system, whereby a master unit controls one or more reactive loads, and whereby at least one follower unit coupled to the master unit responds exactly the same as the master unit to control loads coupled to each follower unit. The master/follower control system according to the seventh embodiment provides for modular flexibility for different sizes of facilities. Figure 9 shows a schematic circuit diagram of a master unit 900. Figure 10 shows a schematic circuit diagram of a follower unit 1000 that is controlled by the master unit 900 of Figure 9. [ooi3i] The seventh embodiment includes a conduction angle phase switching circuit connected in parallel with a reactive load, an AC power source for switching power across the load, and a line switching circuit for enabling the application of AC power to the load through the phase switching circuit. [ooi32] In the seventh embodiment, an ambient light sensor 910 is provided for generating a light control signal indicative of the amount of ambient light present in a particular location. Coupled to the light sensing circuit is a phase angle conduction control circuit, which generates and applies to a control terminal of the phase switching circuit a phase control signal to control the phase angle conduction time of the phase switching circuit, based on the amount of ambient light measured by the light sensing circuit, in order to maintain a substantially constant lighting level. In Figure 9, the microprocessor U3 functions as the phase angle conduction control circuit.

[ooi33] Integrated with the phase angle conduction control circuit is an RC filter circuit which gradually increases the phase angle conduction time switching circuit from zero, or from a predetermined minimum value, to a steady state phase angle conduction time based on the ambient light conditions sensed by the light sensing circuit, after power enabling by the line switching circuit.

[ooi34] Referring to Figure 9, the master unit includes a line switch SWl connected in series with an AC power source between a hot (black) and a neutral (white) power line. Connected in series between the hot and neutral power lines is an reactive load (e.g., fluorescent lamp), and a phase angle control switching device that includes SCRs Ql and Q2 and an opto-isolated triac Ul for driving the SCRs (see discussion with respect to the first embodiment).

[ooi35] Also shown in Figure 9 is the microprocessor U3, which receives a line sync signal from a bridge circuit Dl that is coupled to the hot and neutral lines. Based on the line sync signal, and based on the setting of the potentiometer and switch S Wl , the microprocessor U3 provides control signals to the opto-isolated triac Ul , as well as to follower units coupled to the master unit via pulse width modulated (PWM) signaling.

[ooi36] Figure 10 shows the elements of a follower unit 1000, which receives the PWM control signals from the master unit, and which controls one or more loads connected to the follower unit based on on/off switching of its active element (SCRs Ql, Q2, and opto-isolated triac Ul) via those control signals. [00137] An eighth embodiment of the invention will now be described in detail, with reference to Figures 11-15. The eighth embodiment is directed to a different configuration of the universal control device (UCD) described with respect to other embodiments discussed previously, and whereby the UCD according to the eighth embodiment provides for accurate dimming of Compact Fluorescent Lamps (CFLs) as well as for other types of light fixtures. While the other-described embodiments are capable of dimming CFLs as well as other types of fluorescent fixtures and incandescent fixtures, the UCD according to the eighth embodiment performs this dimming (at least with respect to CFLs) in a more stable and effective manner.

[ooi38] The functions, capabilities, and optional auxiliary hardware defined and mentioned with respect to the other-described embodiments of the UCD apply equally as well to the eighth embodiment. Although the UCD according to the eighth embodiment makes use of the Safety Ground for synchronization purposes only, the UCD can still be considered a "2- Wire" dimmer, in that it can be installed into any standard wall outlet without benefit of the Neutral wire. [ooi39] The UCD according to the eighth embodiment allows more stable and deeper dimming of all types of loads including Magnetic Ballast Fluorescent fixtures, Electronic Ballast Fluorescent fixtures, Compact Fluorescent Lamps, halogen, as well as all types of incandescent lamps, or a motor for a fan or other device. Depending on the application, it may be the case that fluorescent tube life might be reduced somewhat, although there is no concrete evidence of this being the case. Specifically, the UCD according to the eighth embodiment implements a more robust line synchronization circuit, a higher performance processor, and a simpler but more stable dimming algorithm, as compared to the previously- described embodiments of the UCD. [00140] Additionally, software has been written by the inventors for the processor utilized in the UCD according to the eighth embodiment in the standard "C" language. The following paragraphs will describe in detail the differences, advantages, and performance enhancements that are realized by the UCD according to the eighth embodiment.

[ooi4i] CFLs can be dimmed by phase control methods, but success depends on several factors which influence the design of the UCD dimmer. Specifically, the inventors of this application have realized that both voltage synchronization and precise phase timing are important to a stable, flicker free output. Undimmed CFL lamps draw all their current for the entire AC half cycle during the peaks of the AC voltage waveform. This is because they implement a bridge rectifier input charging a capacitor filter to realize DC power, as seen by the bridge rectifier 1310 and the capacitor 1320 in Figure 13. The capacitor 1320 charges only when the AC input is of higher voltage than the capacitor 1320. This occurs only near the peak extremes of the AC input voltage. For CFLs, the period of this current flow is approximately 0.5 mS, and varies with the CFL design itself. Because of this, there is zero dimming of a CFL if the switch element is triggered ON during the first 90° of the AC waveform. In the UCD 1200 according to a preferred implementation of the eighth embodiment, CFL dimming occurs during the 90° to 180° period for the first half of the AC Cycle, and similarly during the 270° to 360° period for the second half of the AC Cycle. Other periods of dimming may be envisioned by one skilled in the art, while remaining within the scope of the invention. [00142] Realistically, a typical CFL lamp will extinguish with a phase delay approaching 160° and 340°. To achieve this degree of phase delay requires an accurate synchronization signal derived from the Line voltage waveform. The previously-described UCD embodiments rely on a noisy and relatively unstable current zero crossing for synchronization. This noisy and relatively unstable sync signal may result at times in some amount of undesirable lamp intensity flickering. The UCD 1200 according to the eighth embodiment facilitates accurate synchronization, and increased phase delay timing that is compatible with flicker free dimming of CFL lamps and other types of fluorescent and incandescent lamps. Referring to Figure 15, the UCD 1200 according to the eighth embodiment utilizes SCRs Ql, Q2, and Opto-Triac Ul, for the solid-state switch unit of the UCD 1200, similar to the other-described embodiments (e.g., see Figure 4). Alternatively, MOSFETs may be utilized for one or more components of the solid-state switch unit.

[00U3] A processor U2 is connected to the Opt-Triac Ul via the OUTPUT line of the processor U2, with a resistor R2 provided therebetween. Similar to the other- described embodiments of the UCD, a potentiometer and switch unit R7 is utilized to set the dimming amount. A potentiometer R8 is also connected to the processor U2, and is used to set the maximum dim level. The potentiometer R8 is a 10KΩ potentiometer and the potentiometer and switch unit R7 also has a resistance of 10KΩ in the preferred implementation of the UCD 1200 according to the eighth embodiment. [ooi44] Also shown in Figure 15 are resistors RlA, RlB, RlC, capacitors Cl, C2, and diodes Dl, D2, which are coupled to the RED line (Load Output). In Figure 15, the resistors R3A, R3B together serve the same function as the resistor Rl shown in Figure 11, and are connected in series to the GREEN safety ground line. [ooi45] One problem with the "current zero crossing" synchronization circuits of the previously-described embodiments of the UCD is that the synchronization circuit implementation is not exceedingly stable with different loads, whereby the synchronization could drift as the load is dimmed. The effect of this can be some slight flickering with certain types of loads, as well as relatively non-linear dimming as the potentiometer is rotated. Also, a load such as a CFL and most electronic ballast devices cannot be adequately synchronized using this type of circuit. As a result, software is utilized for the processor U2 in these earlier- described UCD embodiments in order to maintain a desired level of line synchronization for these current zero-crossing synchronization circuits.

Whenever line synchronization is lost, flickering, flashing, or similar unstable load behavior results.

[00146] Referring now to Figure 11, the UCD 1200 according to the eighth embodiment utilizes a different type of synchronization circuit 1100 that makes use of the Green Safety ground to provide the microprocessor with a very stable load- independent synchronization signal. This effectively provides a stable Voltage Waveform synchronization without violating the concept of a 2-Wire dimming device (the previously-described embodiments utilized current waveform synchronization that synchronize on the zero crossings). [ooi47] One well known specification (Underwriter's Laboratory) sets forth a maximum safety ground current of 500 uA (Micro Amperes) for a wall switch, which may be sourced or sunk by the Green Safety ground. This Green Safety ground is present in every wall switch box. By using a series resistor Rl, such as a 10 megaohm resistor, between the synchronization pin SYNC_INPUT of the microprocessor and the Safety ground, an RMS safety ground current of less than a predetermined amount (e.g., less than 12 uA RMS) is achieved. [ooi48] An important concept in understanding how this works is realizing that the +5 V supply BTl is actually the same as the "Hot" power input (Black Wire). The +5 V power supply BTl actually pulls down the ground through resistor Rl, diode Dl and diode D2 through the load when it is OFF. Battery BTl with its positive terminal tied to the hot side of the AC line represents the UCD circuit power supply for the UCD 1200 according to the eighth embodiment. [00149] In a preferred implementation, the microprocessor synchronization pin SYNC_INPUT (which is initialized as an edge triggered interrupt input) has internal protection diodes Dl₅ D2, which clamp the Sync Input signal to always be between VDD (e.g., +5V) and local ground (LOCAL RETURN). As the Black source wire rises above the Safety ground potential during 0 to 180 degrees, current flows up through protection diode D2, then through resistor Rl, thereby keeping the sync input at local ground potential. Alternatively, as the Black source wire drops below the Safety ground potential during 180 to 360 degrees, current flows down through the Safety ground, through resistor Rl then up through protection diode Dl, keeping the Sync Input at VDD potential. The result is a very clean square wave with very little noise, practically zero phase delay, and nearly totally independent of load type or dimming level.

[00150] The synchronization circuit 1100 utilized in the UCD 1200 according to the eighth embodiment provides a stable synchronization without benefit of the neutral wire. One small side effect as compared to the previously-described embodiments of the UCD is that no longer are the two dimmer wires interchangeable, since the +5 V circuit is tied to the AC Hot Black wire. This is of no major consequence, since the UCD 1200 according to the eighth embodiment still installs in series with the load (see Figure 13, for example), just that the proper polarity of the UCD dimmer needs to be observed for proper operation of the UCD 1200 according to the eighth embodiment.

[ooi5i] Figure 12 shows a waveform obtained by the synchronization circuit 1100, whereby a synchronization signal 1230 that is synched relative to the +5 V power supply is obtained, and whereby the synchronization signal 1230 is high (e.g., logic "1" state, or +5 volts) when the Hot AC Line Input is between 180 and 360 degrees. Synchronization is performed on the falling edge of the synchronization signal in the preferred implementation of the eighth embodiment. tooi52] The previously-described embodiments of the UCD utilize an 8 pin PIC12C672 microprocessor device in the preferred implementation of those embodiments. In the preferred implementation of the UCD 1200 according to the eighth embodiment, a higher performance PIC16C771 microprocessor device is used, as shown by element U2 in Figure 15. One problem with the PIC12C672 is it only had an 8-bit timer, and requires software to sample the timer to initiate switching events. Although this works fairly well in practice, very tight and efficient assembly language software timing is important for stable load control. Also, an 8-bit timer provides only limited dimming resolution (approximately 32 microsecond resolution), whereby discrete dimming steps could be seen by the human eye as the load is dimmed for the previously-described embodiments of the UCD, which is undesirable.

[ooi53] Utilization of a higher-performance microprocessor that implements a more robust (e.g., a 16-bit) timer with a compare interrupt (CCP) function, as in the UCD 1200 according to the preferred implementation of the eighth embodiment, relieves the processor of any critical software execution timing. This hardware interrupt function allows compiled "C" language to be used, thereby greatly simplifying integration and test of the UCD 1200. Additionally, timing resolution is around 2 microseconds for the UCD 1200 according to the eighth embodiment, which provides visually imperceptible dimming steps. Also, the PIC16C771 microprocessor has an internal 12 Bit Analog to Digital converter for increased resolution when reading the Dim and Maximum Dim potentiometers. The result of the use of this stronger processor is a more accurate and stable dimming with less flickering. [ooi54] The UCD 1200 according to the eighth embodiment also utilizes a simpler, more stable dimming algorithm as compared to the previously-described embodiments of the UCD. In the earlier-described embodiments, the synchronization signal is triggered when the series switch element (Triac or Back to Back SCRs) turns OFF. The switch turns OFF when the current through the device reaches Zero. This signal is noisy when controlling inductive loads, requiring the addition of an RC filter network to remove that noise. Depending on the reactive characteristics of the load, and the commanded dim level, this sync signal moves several milliseconds relative to the Voltage zero crossing. Accurate control of the Load is somewhat difficult, requiring complex software to reject spurious noise which occasionally gets through the filter, as well as limiting the amount of dimming possible so that all types of loads could be controlled to some degree without miss-triggering of the switch element. Any miss-trigger or shift in the trigger timing causes the sync signal to temporarily move, resulting in unacceptable flickering.

[ooi55] In conjunction with the synchronization circuit 1100 described earlier and shown in Figure 11, the synchronization algorithm as utilized in the UCD 1200 according to the eighth embodiment becomes simpler, and as a byproduct, more stable. Additionally, switch element triggering can be accurately delayed much longer into the AC waveform providing additional dimming to those devices that can make use of it (e.g., CFLs and Incandescent Lamps) without sync movement, or flickering issues. The synchronization algorithm utilized in the UCD 1200 according to the eighth embodiment is relatively simple in that it simply uses the processor 16-bit Timer to initiate events during the AC cycle. This Timer is effectively synchronized at each rising zero crossing of the AC Voltage waveform by the synchronization circuit 1100 explained earlier. During testing, it was also found that very quick changes to the commanded load dim level could cause the timer compare function (CCPl) to miss incremental interrupt events. To eliminate this problem, a slow dim function can be added, whereby phase timing is moved slowly in response to any speed DimLevel potentiometer adjustment. Further explanation of the synchronization control software will explain this point in more detail.

[ooi56] Along with the upgraded processor, a "C" complier is utilized in the preferred implementation of the eighth embodiment to allow a much simpler but effective program to be written in the "C" computer language.

[ooi57] Figure 13 shows the connection of the UCD 1200 according to the eighth embodiment, from the AC line through the input circuit of the CFL (not shown), whereby the input circuit of the CFL includes a bridge rectifier 1310 and a capacitor 1320, which are connected to a CFL power supply. Higher voltages at the CFL power supply location provide higher lamp intensities, while lower voltages result in lower lamp intensities.

[ooi58] Figure 14 illustrates switching element trigger timing and subsequent CFL current at both high (waveforms designated by "B") and low lamp intensities (waveforms designated by "C"). As is shown in the second and third waveforms in Figure 14, triggering that occurs just prior to the peak (90°) of the AC line voltage (waveform designed by "A") results in both higher level current spikes being transferred to the CFL with a higher applied voltage. This timing results in the highest applied power and resulting lamp intensity. If the switching element trigger is delayed further relative to the applied AC voltage waveform as shown in the fourth and fifth waveforms in Figure 14, the applied voltage is much lower, which results in lower level current spikes, lower applied power, and lower resulting lamp intensity.

[ooi59] The code utilized for the synchronization algorithm according to the preferred implementation of the UCD 1200 according to the eighth embodiment is described below in detail. Declarations: The processor specific header file (16C771.H) is included. Specific register addresses, Bit definitions etc. are found in this file. Then the processor is declared having a 12 bit Analog to digital converter. The Fuses line, declares that the internal 4 MHz clock will be used (INTRC), the internal watchdog timer will not be used (NOWDT), that the security bit protecting against readout of the internally stored code is enabled (PROTECT), that the low voltage brown out detect circuit is enabled (BROWNOUT), the 200 mS power up timer is enabled (PUT), and that the brown out reset voltage is set to 2.7 Volts (BORV27).

Parameter Definitions:

Intwindow: This parameter defines a sync window either side of the synchronization falling edge to be 110 two uS counts (total of 220 uS). This means that the sync interrupt is reset then enabled 220 uS before the expected edge, and if the sync does not occur with 220 uS after the edge, it will freewheel using the calculated sync timing.

TriaconTime: This is the period of time (30 uS) which the OptoTriac device is triggered on.

The OptoTriac will trigger/latch itself as well as the Switch Element ON if a load current of greater than 100 uA flows within this period of time.

Warmtime: This is the number of line syncs that the load will be held to full intensity. For this application 1800 line syncs (16.667 mS each) results in a load warmup time of 30 seconds.

Phaseadv: This parameter is used to advance phase timing to achieve additional load/lamp intensity. This value is determined empirically through test. Define ON/OFF: Defines the State of the output port to turn ON/OFF the Opto-Triac.

Prototypes:

Just a list of functions that occur in the program. Functions must be declared before they appear in the code.

RAM Definitions: Note that Integers define an 8 bit value, and Long defines a 16 bit value.

Phaseswitch: A counter which switches based on which Compare Interrupt (CCPl) occurred.

Status: Four Bits are used, and defined in Bit Definitions

Temp: Short Term 16 bit temporary register

Temp 1 : Short Term 16 bit temporary register Phasecntr: This counter counts full AC cycles, and is used to determine AC waveform timing, as well as the warmup period. Cyclecntr: This is a value representing the number of 2 uSec counts in one AC Cycle. For a 60

Hz system, the resulting value of this counter is nominally 0x208D.

Quartcntr: A value representing 90° of the AC Cycle. Nominally: 0x1869

Halfcntr: A value representing 180° of the AC Cycle. Nominally: 0x1046 Thrqrtcntr: A value representing 270° of the AC Cycle. Nominally: 0x823

Zerocntr: At each Sync edge interrupt the current 16 bit Timer 1 count is stored here

And used to offset all phase timing for the current AC Cycle.

Dimcnt: This is where the calculated phase delay is Stored. Dimcnt is in 2 uSec units.

Maxdim: This is where the calculated phase delay (from reading the MaxDimPot analog value) is stored. MaxDim is in 2 uSec units.

Targetdim: To implement slow changes to the phase delay, the DimLvl pot analog is used as the basis of this targetdim value. The actual dimcnt value is incrementally adjusted to match this

Targetdim value. In this fashion the slow dim function is implemented.

Bit Definitions: Syncdun: Flag which indicates that the Line Frequency timing calcs are completed

Warmdun: Flag indicating that the Warmup time has passed.

Domaxdim: Flag set in the CCPl interrupt to kick off the Maxdim A/D conversion

Dodimlvl: Flag set in the CCPl interrupt to kick off the Dimlvl A/D conversion

Intinput: Defines that Registerδ, Bit 0 is the Edge Triggered Sync Input Address Phase_on: Defines that Registerό, Bit 1 controls the OptoTriac LED

Phase_tri: Defines that Register 86, Bit 1, controls the Tri-State function of the Phase_on output pin.

#INT_CCP1: Compare Interrupt

This interrupt service routine provides the timer 1 based control of the Switch element. The value of phaseswitch is used to determine which event will be serviced during this interrupt. For each AC cycle, there will be four CCPl compare interrupts. Phaseswitch cycles from 1 through

4 for each AC cycle. During each event, the calculated time of the next event is loaded into the

CCP_1 register. When the just loaded CCP_1 value is equal to the free running Timer 1 value, the next CCP_1 interrupt is generated. In the manner, very accurate phase timing can be implemented with minimal regard to processor execution speed. The zerocntr value is loaded at each sync interrupt and used to offset the loaded timerl count during the CCP_1 interrupt. Timer 1 is allowed to free run, thereby eliminating any latency issues associated with resetting or reloading Timer 1 with specific values.

Phaseswitch =1: This CCP_1 interrupt occurs at the sync interrupt plus the intwindow value. Nominally this will be at 220 uS after the sync zero crossing. This signals the end of the synchronization window. If the sync interrupt had occurred, zerocnt would have been adjusted in the sync interrupt routine. If it didn't, then we're free wheeling from the last zerocntr + cyclecntr values. At this time, the sync interrupt is disabled. Additionally, the domaxdim flag is set, whereby the maxdim analog conversion will be updated during processor idle time. The next CCP_1 compare value is loaded. Note that the zerocntr value increments nominally by on cyclecntr value each AC cycle, the quartcntr value is static, and the dimcnt value represents the calculated phase delay which corresponds to the commanded load dim level. Phaseswitch -2: This CCP_1 interrupt occurs at zerocntr + quartcntr + dimcnt. This is time that the Switch element is triggered ON for a period of triacontime in uSec. The Switch element remains latched ON until the current through it drops to zero. This will occur sometime after the next AC voltage waveform zero crossing, depending on the reactive characteristics of the load. The next CCP l compare value is loaded.

Phaseswitch =3: This CCP_1 interrupt occurs at zerocntr + thrqrtcntr + dimcnt. Functionally this event is identical to Phaseswitch =2, except it occurs exactly one half cycle later. Additionally, the dodimlvl flag is set, whereby the dimlvl analog conversion will be updated during processor idle time.

Phaseswitch =4: This CCP_1 interrupt occurs at zerocntr + cyclecntr - intwindow. Effectively, this is the next sync interrupt - intwindow, or nominally at 16.667 mS - 220 uSec from the last sync interrupt. This is the start of the interrupt window. During this event, any sync interrupt flags are cleared, and the sync interrupt is enabled. Also the value of phaseswitch is cleared. Upon Exit, the value of phase switch is incremented. In addition to during the Phaseswitch =4 event, the value of phaseswitch is also cleared during each sync interrupt. In this manner, if a sync interrupt is missed, the value of phaseswitch will be maintained while syncs are free wheeling. #INT_EXT: Line Synchronization External Interrupt This interrupt service routine calculates line frequency parameters, as well as updates the zerocntr, phasecntr, and phaseswitch values. This interrupt is initialized to occur at each falling edge of the sync input pin. During initialization, 20 sync interrupts are allowed to pass while things stabilize. At the 20^th sync, Timerl is set to zero. When the 26^th sync occurs, Timerl has free run from a count of zero for 6 full AC cycles. Timer 1 is incremented at a 500 Khz rate. The cyclecntr, halfcntr, quartcntr, and thrqrtcntr timer values are calculated from the time it took for 6 complete AC cycles to occur. Measuring/calculating timing in this fashion reduces the need for an accurate microprocessor oscillator, plus provides compatibility with both 50 and 60 Hz line frequencies. After these line timing parameters are calculated, the Timerl compare interrupt (CCP_1) is enabled to occur at the current time plus 30 counts. Finally, the syncdun flag is set indicating that the sync timing calculations have been completed. * Main(), Reset(): Main and Reset Functions

Since this program is entirely interrupt driven, the Main () function is quite simple. Additionally, if additional functions need to be added, there is ample processor time available between interrupts.

Reset(): First the Reset function is called, whereby the internal microprocessor peripherals are initialized, memory variables are initialized, and the external sync interrupt is enabled. Then the program waits for completion of the sync timing calculations. When this is complete, an adjustment of -phaseadv counts (Nominally 200) is made to the quartcntr and thrqrtcntr values. These two values are only used to trigger the Switch element ON. This was an easy way to increase the effective phase timing available (increased load/lamp intensity) while maintaining line frequency independence, and preserving the circuit power supply.

After returning from Reset, Main compares the value of phasecntr to see if it's reached the warmtime count. If so it sets the warmdun flag, allowing normal load dimming to occur. Next a very simple Flag based task switcher calls the rdmaxdim(), and rddimlvl() functions.

• Rddimlvl(): Read the DimLvl potentiometer, Slow Dim Implementation

Although A/D conversions are 12 bit values, only the most significant 10 bits are used. The PIC16C771 stores these 12 bit conversions left justified in a 16 bit variable to maintain code compatibility with microprocessors with 8 bit A/D converters. To fix this, the stored 16 bit value must be right shifted by 6, then ANDed with 0x3FF to result in a 10 bit result. A 10 bit result has a range of 0 to 1024 decimal. Because perceived dimming is not linear with phase shift , a simple one point correction is implemented. If the 10 bit dimlvl is less than 300, then it is multiplied by two. This will span a count range of zero to 1198 counts, while giving a slope of two. If it is greater or equal to 600, 600 is Added to it. This will extend the range from 1200 to 1624 with a slope of One. The total extent of the calculated value is 0 to 1198 with a slope of Two, and 1200 to 1624 with a slope of One. When used to control the Timer 1 Compare interrupt which has a resolution of 2 uSec, the total timing range is 0 to 3.248 mSec. Targetdim is loaded with this calculated value each AC cycle.

Slowdimming: Each AC cycle, the variable dimcnt is incremented or decremented until it matches the targetdim value. In this manner dimcnt always chases the value of targetdim. To allow dimcnt to always catch targetdim in a reasonable period of time, a two speed step is implemented. The code provides a step of either 1 or 4 each AC cycle, depending on how far away dimcnt is from targetdim. In the event of a full range step, dimcnt will catch target dim in 1624 counts /4 counts per step * 16.667 mSec per step or 6.8 Seconds. This would be the absolute worst case time.

^• Rdmaxdim(): Limit the maximum dim level

Because each load behaves differently when dimmed, a Maxdim potentiometer is provided which limits the total phase delay that dimcnt can achieve. Similar to when reading the dimlvl potentiometer, a 10 bit conversion is calculated. In this case, an offset of 500 counts is added to it. This corrected maxdim value is used in rddimlvl() to clamp dimcnt. [00160] The C-language software utilized by the processor U2 in the UCD 1200 according to a preferred implementation of the eighth embodiment, is provided below for sake of completeness.

SOFTWARE IN "C":

// Initial Code Date: 8/14/04//For PIC16C771 Chip Sept 12,2004

// UCD-CFL This code uses the UCD hardware and the 16C771 Microprocessor to dim CFL Lamps

// The Safety Ground (Green Wire) is used for Line Synchronization // UL indicates a Maximum of 500 micro amps may be dumped into the Safety Ground

// A minimum of 9.3 Megohms to the Safety Ground will Guarantee that less than

// 20 Micro Amps Peak will flow. RMS will be less than 15 Micro Amps

// Dual Slope dim level delay timing is implemented to "linearize" the Pot.

// A/D conversions are the Ratio of +5V supply for supply voltage rejection. // All calculations are line frequency independant, providing compatibility to 50 or 60 Hz systems.

// Tools/Compiler Option Line: C:\Program Files\PICC\Ccsc +FM +DS +DC

// The Program

//Declarations

#include <16c771.h> #device PIC16C771 ADC=12 //12 Bit Conversions

#fuses !NTRC,NOWDT,PROTECT,BROWNOUT,PUT,BORV27

#use delay(clock=4000000) //4000000 4MHz Clock for Delays //Parameter Definitions // 2 uSec Timer Counts #define intwindow 150 // intwindow: 50 is Boarderline: Timer count Window for Interrupt #define triacontime 30 // triacontime: Time the Triac is Held ON in Micro Seconds #define warmtime 1800 // warmtime: 1800 = 30 Seconds. Num Syncs to warm up Load #define mindim 400 #define strtdim 1900 #define ON 0x00 //Turning the Load ON requires a 0 to be written to the output pori #define OFF 0x01 //Turning the Load OFF requires a 1 to be written to the output port. //Prototypes void comp_int(); void sync(); void reset(); void rddimlvIO; void rdmaxdim(); //RAM Definitions int phaseswitch; //Phase Switch Register int status; //Status Register long temp; //Temp Register long tempi ; //Temp Register long diminc; //Increment each loop to change dimcnt long phasecntr; //Counts AC cycles, Used for timing and warmup long cyclecntr; //Timer Counts for One full Cycle long eightcntr; //Timer Counts for 45 degrees long halfcntr; //Timer Counts for 180 degrees long fivδcntr; //Timer Counts for 225 degrees long zerocntr; //Timer Count at Sync Interrupt: Zero Count Reference long dimcnt; //Calculated phase delay from dimlevel pot long maxdim; // maxdim: Maximum counts to delay trigger, Higher Numbers go dimmer long targetdim; //targetdim: Accumulator for Dim for slow dim

//Bit Definitions

#bit syncdun =status.O //Sync Calculations Done

#bit warmdun =status.1 //30 Sec. Warm UP done

#bit domaxdim =status.2 //Set during each sync interrupt, Kicks off A/D conversions

#bit dodimlvl =status.3 //Set during each sync interrupt, Kicks off A/D conversions

#bit intinput =0x06.0 //Sync Input

#bit phase_on =0x06.1 //RB 1 Load ON =0, Load OFF =1

#bit phasejri =0x86.1 //RB 1 Tristate is 0 = Output

#INT_CCP1 //Timer! = Compare Value Interrupt comp_int() { switch (phaseswitch){ case 0x01: { // Zero Crossing ÷intwindow

DISABLE_INTERRUPTS(INT_EXT); temp = (zerocntr ÷eightcntr +dimcnt); //Dimcnt is the Variable Dim Count

CCP_1 = temp; //CCP 1: Deadi domaxdim =1; break;

} case 0x02: { //1st Half of Cycle: Time to Trigger ON Triac phase_on =ON; delay_us(triacontime); //Software Timer here, Not Critical phase_on =OFF; //Triac Turns OFF when Load current is Zero temp = (zerocntr +fivδcntr +dimcnt);

CCP_1 = temp; //Triac Latched ON, Turns OFF when Load current is Zero break;

} case 0x03: { //2nd Half of Cycle: Time to Trigger ON Triac phase_on =ON; //Software Timer here, Not Critical delay_us(triacontime); //Triac Latched ON Turns OFF when Load current is Zero phase_on =OFF; //Triac Turns OFF when Load current is Zero temp = (zerocntr ÷cyclecntr -intwindow); CCP_1 =temp; dodimlvl =1; break;

} case 0x04: { //Set Up for Next Line Sync Interrupt zerocntr +=cyclecntr; phaseswitch =0; //Incremented to 1 before lnt Exit

CCP_1 =(zerocntr ÷intwindow); #asm bcf OxOb, 1 //Clear External Interrupt Flag #endasm

ENABLE_INTERRUPTS(INT_EXT); break;

}

} phaseswitch +=1; //phaseswitch rolls 1 to 4

} //End comp_int()

#INT_EXT //External Interrupt on CPUSYNC sync() { //Falling Edge zerocntr =GET_TIMER1(); //Get counter at Interrupt, Timer Resolution is 2.000 uSec phasecntr +=1 ; //TMR1 overflows at Cycles, Stop it at 6 phaseswitch =1 ; if (syncdun ==0) { if (phasecntr >= 26) { //First Time Through cyclecntr = zerocntr /6; //Cyclecntr is 1/6th of 6 cycles or 1 Cycle halfcntr = zerocntr /12; //Halfcntr is 180 degrees eightcntr = zerocntr /48; fivδcntr = halfcntr ÷eightcntr;

CCP_1 =(zerocntr +50); //Kick first CCP Interrupt

ENABLE_INTERRUPTS(INT_CCP1); //CCP ints are enabled on after sync timing is done syncdun =1 ; //Done with Sync Timing

} if (phasecntr == 20) //Let a few Interrupts in B4 Clearing Counter

SET TIMER1 (0x0000); //Only done at Startup, Otherwise Rolls }

} //End syncO main() { reset(); while (1) { //Warm Up time, Load is ON if ((phasecntr >= warmtime) && (warmdun ==0)) //1800 = 30 Seconds warmdun =1; while (domaxdim ==0); //Wait for next sync rdmaxdim(); while (dodimlvl ==0); //Wait for next sync rddimlvIO; }

} //End main() void reset() { //Reset Function Initialize RAM phase_on -OFF; //Turn off Load phasecntr =0; cyclecntr =O; targetdim =0; phaseswitch =1 ; //Phase Switch Register, Rolls 1 to 4, no zero halfcntr =0; syncdun =0; warmdun =0; diminc =4; //During Startup, Dim Level Changes are Very Slow phasejri =0; //TrisB 0x86.1: 0 = Output, 1 is Input dimcnt =strtdim; maxdim =strtdim; targetdim = mindim; //Count from Startdim to Min Dim during Warmup rdmaxdim();

SETUP_ADC(ADC_CLOCK_DlV_32); SETUP_ADC_PORTS(sAN0 | sAN3);

SETUP_TIMER_1(T1_INTERNAL | T1_DIV_BY_2); //Timeri Fosc/4 = 25OnS, Free Run delay_ms(400); //Wait Additional 400 mS for Power to Stabilize

SETUP_CCP1 (CCP_COMPARE_INT); EXT_INT_EDGE(H_TO_L); //Falling Edge of Line Sync

#asm bcf OxOb, 1 //Clear External Interrupt Flag #endasm

ENABLE_INTERRUPTS(INT_EXT); ENABLEJNTERRUPTS(GLOBAL); while(syncdun ==0); //Wait until Waveform Timing is Complete

} //End reset() //Load is OFF upon exit void rddimlvIO {

SET_ADC_CHANNEL(1); //Set Mux for Dim Level delay_us(10); dodimlvl =0; //Dual Slope Dimming to Linearize Pot And if (warmdun ==1) { //Slow down Dimming targetdim = (((READ_ADC() »4) &0x0fff)+mindim); //11 Bit Number 2048 Max if (targetdim > maxdim) targetdim =maxdim; diminc =10;

} if (dimcnt >(targetdim +diminc)) dimcnt -=diminc; if (dimcnt <(targetdim -diminc)) dimcnt +=diminc; if (dimcnt >targetdim) dimcnt -=1; if (dimcnt <targetdim) dimcnt +=1 ;

} //End rddimMO void rdmaxdim() {

SET_ADC_CHANNEL(3); //Set Mux for Max Dim Level delay_us(10); domaxdim =0; maxdim = (((READ_ADC() »5) &0x07ff)+700); //11 Bit Number } //End rdmaxdim() tooi6i] The table below provides a list of components for the UCD 1200 according to a preferred implementation of the eighth embodiment. Table: List of Components for UCD 1200

Item Quantity Reference Part Vendor Part Number

1 1 C1 47OuF 10V Digi-Key P5126-ND

2 1 C2 0.47uF 630V Digi-Key EF6474-ND

3 1 C3 Not Installed

4 1 C4 .047 50V Digi-Key P4965-ND

5 1 D1 5.6Vz 1W Digi-Key 1N4734DITR-ND

6 1 D2 1N4001 Digi-Key 1N4001 GICT-ND

7 1 D3 5.1Vz 1/2W Digi-Key 1N5231 BDITR-ND

8 2 J2.J1 Thru Holes

9 1 Q1 Q6016LH3 Digi-Key Q6016LH3 ND

10 1 R1 620 1/2W Digi-Key 620H-ND

11 1 R2 180 1/2W Digi-Key 180H-ND

12 1 R3 51K 1/4W Digi-Key 51 KQBK-ND

13 1 R4 51K 1/8W Digi-Key 51 KEBK-ND

14 1 R5 100K 1/8W Digi-Key 100KEBK-ND

15 2 R6.R9 4.7K 1/8W Digi-Key 4.7KEBK-ND

16 1 R7 620 1/8W Digi-Key 620EBK-ND

17 1 R8 Not Installed

* 18 1 SWPOT1 10K POT + Switch CTS 450S201

19 1 U1 MOC3022 Digi-Key 160-1376-5-ND

20 1 U2 P1C16C774 Digi-Key IC16C774 -ND

21 1 F1 EMl Filter

22 1 W Wire 8 -12 " Long (CON1) Belden 16 AWG Color GRN (w/label "Safety

23 1 W Wire 8 -12 "' Long (CON1) Belden 16AWGColor BLK (w/label "AC Line"

24 1 W Wire 8 -12 " Long (CON1) Belden 16AWGColor RED (w/label "Load")

* NOTE : (Except Molded Shaft Construction)

[00162] Different embodiments of the present invention ^'. have been described according to the present invention. Many modifications and variations may be made to the techniques and structures described and illustrated herein without departing from the spirit and scope of the invention. Accordingly, it should be understood that the apparatuses described herein are illustrative only and are not limiting upon the scope of the invention. With the use of an energy savings device according to an embodiment of the invention, it is possible to achieve a 50% or more energy savings, while not adversely affecting the perceived amount of light by users.

[00163] Also, the above-described embodiments of the present invention are capable of providing dimming for electronic ballast fluorescent fixtures, using the same electronics and software as those described earlier with respect to magnetic ballast fluorescent fixtures. Tests performed by the inventors showed a dimming capability for several different types of electronic ballast fluorescent fixtures, without any noticeable flickering. Therefore, an apparatus and method according to different embodiments of the present invention can be used to control resistive, inductive, and/or capacitive loads.

Claims

WHAT IS CLAIMED IS:

1. An energy savings device for an inductive, resistive or capacitive load that is powered by an input AC voltage waveform, comprising: a setting unit configured to allow a user to set a desired power operating level for the load; a processor configured to receive a signal from the setting unit indicative of the desired power operating level for the load, to determine a phase delay to be provided to an output AC voltage waveform that is to be provided to the load, and to output a control signal as a result thereof; and a synchronization circuit for synchronizing to a Safety Ground line, the synchronization circuit providing a synchronization signal that is used in construction of the control signal by the processor.

2. The energy savings device according to claim 1, wherein the active element comprises: a first SCR having an anode terminal coupled to the line and having a cathode terminal coupled to the load; and a second SCR coupled in parallel to the first SCR, the second SCR having a cathode terminal coupled to the line and having an anode terminal coupled to the load.

3. The energy savings device according to claim 2, further comprising: an opto-isolated triac provided between the processor and the active element, the opto-isolated triac providing the control signal to the first and second SCRs while providing a protection function for the microprocessor.

4. The energy savings device according to claim 1, wherein the load is either a fluorescent light fixture having either an external magnetic ballast or an electronic ballast, or a compact fluorescent light fixture having an internal ballast, or a motor for a fan or other device.

5. The energy savings device according to claim 4, further comprising: a motion detector configured to detect any motion within a particular area, and to provide a motion signal to the processor indicative as to whether or not any motion is detected, wherein the processor is configured to control a dimming level of the fluorescent light fixture based in part on the motion signal.

6. The energy savings device according to claim 1, further comprising: an active element provided between a line that provides the input AC voltage waveform and the load, the active element receiving the control signal and turning off and on at predetermined times in accordance with the control signal, so as to create the output AC voltage waveform from the input AC voltage waveform.

7. An energy savings method for an inductive, resistive, or capacitive load that is powered by an input AC voltage waveform, the method comprising: setting a desired power operating level for the load; receiving, by a processor, a signal indicative of the desired power operating level for the load, and determining a phase delay to be provided to an output AC voltage waveform that is to be provided to the load, and to output a control signal as a result thereof; and synchronizing to a Safety Ground line, the synchronizing step resulting in a synchronization signal that is used in construction of the control signal by the processor.

8. The energy savings method according to claim 7, wherein the active element comprises: a first SCR having an anode terminal coupled to the line and having a cathode terminal coupled to the load; and a second SCR coupled in parallel to the first SCR, the second SCR having a cathode terminal coupled to the line and having an anode terminal coupled to the load, or a motor for a fan or other device.

9. The energy savings method according to claim 7, wherein the load is either a fluorescent light fixture with an external magnetic ballast or an electronic ballast, or a compact fluorescent light fixture with an internal ballast.

10. The energy savings method according to claim 9, further comprising: detecting any motion within a particular area, and providing a motion signal to the processor indicative as to whether or not any motion is detected; and controlling a dimming level of the fluorescent light fixture based in part on the motion signal.

11. The energy savings method according to claim 7, further comprising: receiving the control signal, and, in response thereto, turning an active element off and on at predetermined times in accordance with the control signal, so as to create the output AC voltage waveform from the input AC voltage waveform, wherein the active element is disposed between a line carrying the input AC voltage waveform and the load.

12. A computer program product being executed by a processor and which provides an energy savings capability for an inductive, resistive or capacitive load that is powered by an input AC voltage waveform, the computer program product comprising: first computer code configured to set a desired power operating level for the load; second computer code configured to receive a setting signal output from the first computer code that is indicative of the desired power operating level for the load, the second computer code further configured to determine a phase delay to be provided to an output AC voltage waveform that is to be provided to the load, and to output a control signal as a result thereof; and third computer code configured to synchronize to a Safety Ground line, the third computer code providing a synchronization signal that is used in construction of the control signal.

13. The computer program product according to claim 12, wherein the active element comprises: a first SCR having an anode terminal coupled to the line and having a cathode terminal coupled to the load; and a second SCR coupled in parallel to the first SCR, the second SCR having a cathode terminal coupled to the line and having an anode terminal coupled to the load.

14. The computer program product according to claim 12, further comprising: an opto-isolated triac provided between the processor and the active element, the opto-isolated triac providing the control signal to the first and second SCRs while providing a protection function for the processor which executes the first, second, and third computer codes.

15. The computer program product according to claim 12, wherein the load is either a fluorescent light fixture with an external magnetic ballast or an electronic ballast, or a compact fluorescent light fixture with an internal ballast, or a motor for a fan or other device.

16. The computer program product according to claim 12, further comprising: fourth computer code configured to detect any motion within a particular area, and to provide a motion signal to the processor indicative as to whether or not any motion is detected, wherein the processor is configured to control a dimming level of the fluorescent light fixture based in part on the motion signal.

17. The computer program product according to claim 12, further comprising: fourth computer code configured to provide a control signal to an active element provided between a line that provides the input AC voltage waveform and the load, the active element receiving the control signal and turning off and on at predetermined times in accordance with the control signal, so as to create the output AC voltage waveform from the input AC voltage waveform, wherein the control signal is provided based on the phase delay determined by the second computer code and the setting signal output by the first computer code.

18. An energy savings device for an inductive, resistive or capacitive load that is powered by an input AC voltage waveform, comprising: setting means for allowing a user to set a desired power operating level for the load; processing means for receiving a signal from the setting unit indicative of the desired power operating level for the load, and for determining a phase delay to be provided to an output AC voltage waveform that is to be provided to the load, and to output a control signal as a result thereof; and synchronization means for synchronizing to a Safety Ground line, the synchronization means providing a synchronization signal that is used in construction of the control signal by the processing means. signal conversion means, provided between a line that provides the input AC voltage waveform and the load, for receiving the control signal and turning off and on at predetermined times in accordance with the control signal, so as to create the output AC voltage waveform from the input AC voltage waveform.

19. The energy savings device according to claim 18, wherein the signal conversion means comprises: a first SCR having an anode terminal coupled to the line and having a cathode terminal coupled to the load; and a second SCR coupled in parallel to the first SCR, the second SCR having a cathode terminal coupled to the line and having an anode terminal coupled to the load.

20. The energy savings device according to claim 19, further comprising: isolation means provided between the processing means and the conversion means, the isolation means providing the control signal to the first and second SCRs while providing a protection function for the processing means.

21. The energy savings device according to claim 18, wherein the load is either a fluorescent light fixture having an external magnetic ballast or an electronic ballast, or a compact fluorescent light fixture having an internal ballast, or a motor for a fan or other device.

22. The energy savings device according to claim 21, further comprising: motion detection means for detecting any motion within a particular area, and to provide a motion signal to the processing means indicative as to whether or not any motion is detected, wherein the processing means controls a dimming level of the fluorescent light fixture based in part on the motion signal.

23. The energy savings device according to claim 18, wherein the setting means comprises a rotatable knob provided on a wall.

24. The energy savings device according to claim 18, further comprising: signal conversion means, provided between a line that provides the input AC voltage waveform and the load, for receiving the control signal and turning off and on at predetermined times in accordance with the control signal, so as to create the output AC voltage waveform from the input AC voltage waveform.