Improvements to led dimmer and driver design

Abstract

An LED dimmer having a switch mechanism M1 which enables the switching on and off of a low impedance path from one port of an AC supply V1 to a serial feed capacitor C1, the other end of which is conn

Description

Description History Light Emitting Diodes (LEDs) have been available for general lighting for several years now and they have been shown to be more reliable and cheaper to run than previous types of light bulbs. However the driver circuitiy used in most lamps today is not very efficient and can use over 30% of the power derived from the mains supply. The invention described here brings that quantity down to close to 5%. Today, in the early 2020s, LEDs have electrical to light conversion factors of around 40%. Clearly the hope is that this will improve over time. That conversion efficiency is not the subject of this invention. This invention is concerned with converting the Mains supplied AC to a DC supply suitable for driving the LEDs and the efficiency and reliability of that conversion. Many existing non dimming Driver circuits include a constant current integrated circuit. This is required to maintain a steady current supply to the LEDs, so that they will not appear to flicker to the user. However these do reduce the overall efficiency by around 20%. In addition inductors and resistors in series with the LEDs will also reduce the efficiency so that the net efficiency will be around 65%. Existing drivers which are designed for dimming use fairly complicated circuitry including integrated circuits designed to work with existing Leading or Trailing Edge dimmers. Because the novel driver circuit in this document does not have this complicated circuitiy it does not operate well with Existing dimmers using Leading and Trailing Edge technologies so a new type of dimmer is needed which will work with this new driver circuit. As already mentioned LED dimmers as used today can be put into two categories, Leading edge dimmers and Trailing edge dimmers. The dimmer described in this document surpasses these old designs and can be called a MED (Multi-Edge Dimmer). Referring to Figure 6, the sinusoidal wave representing the voltage between the live and neutral lines of a Mains AC. Leading edge dimmers use a control mechanism for controlling the mains supply to the driver that use a switch-on command at Positions B, switching-off at the zero crossing, while Trailing edge dimmers switch on at the zero crossing and switch off at Positions A. This means that they use the zero crossing of the Mains AC waveform to either switch on or switch off. This new dimmer does not use the zero crossings to switch on or switch off. This new dimmer may be arranged to switch-on and switch off multiple times anywhere during the AC cycle. For example it could be arranged to switch on at Positions B and switch off at Positions A. Furthermore this new Dimmer can also be arranged to switch on at Position C and switch off at Position D in Figure 7, more details will be given later in this document. LEDs are not only much more efficient in terms of converting electrical energy into light than previous technologies, they also are much more reliable. Typical domestic light bulbs using incandescent technology had a lifetime of around 1000 hours while LEDs are often guaranteed over 20,000 hours. This is clearly much better than incandescent but given the high cost of actually replacing bulbs in many situations, I consider that there will be a strong market for the 100,000 hours MTBF lightbulb. Essentially this will mean that the bulb will last 10 years continuously used. This new design of Driver and Dimmer enables the lifetime of the LEDs to be so extended. Description This new Driver design uses fewer electronic components than existing designs. In particular, the complex integrated circuit used to stabilise the current through the LED is not required. This new Driver design is also extremely efficient, capable of converting over 95% of the power from the supply direcdy to the LEDs. Reliability of any electronic device is a function of the temperature in its environment. In many cases this function can be approximated to a halving of the lifetime of the device when the temperature is raised by lOdeg Celsius. Experiments have shown that with a standard lOwatt LED lamp, by eliminating unnecessary components from a LED lamp, the efficiency is raised so that the current and therefore the power consumption is reduced leading to the reduction of the temperature of the body of the LED lamp is reduced by some 20deg Celsius. Furthermore the removal of electronics from the Light bulb removes a number of semiconductor junctions which also increases its reliability so that we can be confident that we have increased the lifetime of the light bulb by at least a factor of 5. If no Dimmer is to be used the new Drivers described here can be connected directly to the mains (see Figure 1), with an example of a schematic shown in Figure 5. If a Dimmer is to used, then it is necessary to consider Figure 2, in which the block diagram shows three distinct areas in an LED lamp control structure using a Driver and a Dimmer. At the right hand side of each of the Figures 1, 2 and 3 the LEDs in one lamp are shown. These are the LEDs which will produce the light. In addition there are two distinct parts of the circuitry described here, the Dimmer circuitiy and the Driver circuitry. In practice the Driver circuitry would usually be housed with the LEDs inside the body of the lamp, while the Dimmer circuitry would be housed separately. If just one LED lamp is to be used, then Figure 2 is the situation. The Dimmer could also be connected to several Driver circuits connected in parallel in several lamps (see Figure 3 and Figure 12). Only 3 lamps are shown in order to simplify the Figures. Any number of lamps could be connected to one Dimmer, the number only limited by the power output capability of the dimmer itself. The Dimmer itself can be divided into 3 main functional blocks as shown in Figure 4. These will described in more detail in this document. Circuit diagram of the Driver (No dimming) When no dimming is required, all that is required is the Driver circuit connected to the LEDs (Figure 1). This can connected directly to the 120 / 230VAC mains supply. In Figure 5, the Circuit Diagram of the driver and a serially connected set of LEDs is shown. The mains supply is shown by VI, where one connection is connected to the Feed Capacitor Cl. Then the other terminal of Cl and the other mains connection are connected to a Full Wave Rectifier consisting of 4 diodes, DI, D2, D3 and D21. This Full Wave Rectifier then rectifies the mains and a DC voltage with ripple appears across the terminals of the Smoothing Capacitor C2. In this example, the Driver may be considered to operats as a current source, the magnitude of the current depending on the capacity of the Feed Capacitor Cl. In this example when Cl is a 2.2uF capacitor, it produces a constant current of around 90mA at 100VDC across the LEDs. A Polypropylene film Capacitor has a negative temperature coefficient and a veiy low Dissipation Factor. These properties make these capacitors ideal for this application (Cl). This is then fed to LEDs D5 to D20. The ripple on the capacitor C2 and across the LEDs will depend on the value of C2. The value of C2 can be modified to the specific requirements of the design. Values as low as lOuF have been shown to produce excellent results and if required a perfectly good output may be seen even without the presence of C2. The determining factors are the properties of the LEDs used. Some LEDs are very sensitive to peak current and if it needs to be carefully controlled then C2 will probably be needed and its value will determine the peak current to average current ratio. See Figure 10 and Figure 11 for simulated examples. It should be noted that the Driver is exactly the same for Non-Dimmed as well for Dimming designs. The only difference is that the Driver is directly attached to the Mains supply when non dimming is required and attached to the Dimmer when dimming is required. It should be noted that no series resistor is used in this circuit. Any series resistance will reduce the efficiency of the circuit. Of course, if regulations require a safety temperature triggered fuse or similar component this would be added. Power Efficiency of the driver Defining the Power Efficiency of the Driver as the Input Power to the LEDs divided by the Output power from the mains supply, then one of the main parameters affecting the Power efficiency is the Dissipation Factor (DF) of the Capacitor Cl. Film Capacitors have a low DF of the order of 0.05% and are readily available in suitable values. The main negative issue with them are their size. If size is critical then multi-layer Ceramic capacitors will do but they do have an DF of 2% which needs to be considered carefully when heat dissipation becomes an issue. With a low Dissipation Factor capacitor, then efficiencies of over 95% are possible. Peak to average current ratio Peak to average ratio of the current through the LEDs is a key parameter for some LEDs. In general, peak to average ratios of more than 3 should be avoided in order to maintain the reliability of the LED. Using the circuit shown in Figure 5, Figure 10 shows the simulated current through the LEDs with Cl=2uF, without capacitor C2 and Figure 11 show the current with C2 =100uF. The average current is around 147 mA. From Figure 10 and Figure 11 it is seen that the instantaneous current through the LED swings from 0mA to over 220mA without C2 and this changes to a minimum of 55mA and a maximum of 175mA with C2=100uF. This demonstrates that the variation between the minimum and maximum current is about halved with the lOOuF capacitor. It should be noted that in each case the peak to average current is less than 3. Driver Analysis Figure 5 depicts a set of LEDs connected in series being powered by a mains supply. In Figure 5 C2 can be seen as a smoothing capacitor to reduce the peaks and fill in the zeroes of the output of the 4 diodes, however due to the behaviour of the human eye, in general this capacitor is not necessary to provide an apparently stable light output from modern LEDs. If we assume a 50Hz mains supply frequency, the impedance of Cl is (l / (2*n*50*Cl)) Assume that the voltage across all the LED load is a 50V DC. Then the current flowing through the LEDs, Lamp, will be approximately Ilamp=(V-50)nns / (l / (300*Cl)) With Vrms=240, from simulation this can be simplified to approximately: Ilamp 233*300*Cl For Cl with a value of luF, this gives a current of 70mA. This shows that this simple circuit, the addition of a Smoothing Capacitor will provide an approximately constant current to the LEDs which is directly proportional to the capacitance of Cl. Losses in the circuit are due to the Dissipation Factor (DF) of Cl and the voltage drop and losses across the diodes DI, D2, D3 and D21. If film capacitors with DFs of less than 0.05 provide total losses of less than 5%. Operation principles when Dimming is required Dimming the LED Good quality dimming is a critical capability for a general purpose lamp and the lamp circuit described above in Figure 5 is not suitable for dimming with existing Leading and Trailing edge dimmers. A new and novel dimmer is required for this new LED driver. This dimmer which is a Multi Edge Dimmer (MED) combines the features of both the Trailing and Leading Edge dimmers and is eminently suitable for this very efficient LED driver. Dimming the LEDs requires that the average current provided to the LED is smoothly altered as well as limiting the maximum current as seen by the LEDs. The vaiying currents need to be effected in such a way so that the light output from the LEDs appears to be constant to the human eye. There are various ways this can be achieved. Using the formula for the current derived in the previous section, it can seen that there are two possibilities which would adjust the current through the LED. The first would be to change the value of the Capacitor Cl in Figure 5, and the second is to change the voltage (V-50)rms. Altering the value of Cl in Figure 5 If the value of Cl is changed to effect a dimming function, keeping (V-50)rms the same, then in a situation where there are multiple lamps controlled by one dimmer, then each lamp will need to have this capability. If each lamp did not have this capability, then the dimmer would need to be designed and built for the number of lamps it will be driving. That would be a huge burden on the manufacture and supply of dimmers. It is much better if every dimmer can be used for any number of lamps up to a specified maximum. Therefore this technique of changing Cl will not be pursued. Altering the value of Vrms The another method to effect dimming is to adjust Vrms and keep Cl constant. With this method, the dimmer as shown in Figure 3 can adjust the value of Vrms for all the connected lamps, so any number of lamps can be controlled by one dimmer up to its maximum load. There are several methods to produce this result and one example is described below: The Dimmer is controlled by an Internal Control unit (See Figure 4), which collects information from the Synchronisation unit so that the control signals which are sent to the Switching unit are time synchronised to the Mains AC cycle. The Internal Control unit then sends signals to the Switching unit to switch the Mains AC on and off to the Driver unit at times determined by the Internal Control unit. Switching on the Mains AC means allowing the Mains AC voltage to be connected to the Driver Unit with low impedance, ideally zero impedance. Switching off means isolating the Mains AC voltage from the Driver with a high impedance, ideally infinite impedance. The Control unit itself is under the control of an External control unit which resides outside the physical dimmer itself which maybe via a wired or a wireless connection. In general the External Control unit will determine the Power output required from the LEDs, and this required output will be converted in the Internal Control Unit into a series of Switch-on and switch-off commands to the Switching Unit. This switching on and switching off will provide controlled sequences of electrical charge synchronised to the Mains AC cycle to be sent to C2 in Figure 9 so that the average voltage and hence the current delivered to the LEDs can be controlled. The peak to average current being delivered to the LEDs need to be as close to 1 as possible to minimise the variation of light being produced by the LEDs. This is also important because LEDs will be damaged or the life of the LED can be reduced if too high peak currents are used. In addition the average current supplied over a mains cycle should be maintained as possible as possible so as to reduce flicker and light output variation. The human eye can tolerate a significant amount of light variation over a short time (10ms or so) without it being an issue but variations longer than that are unpleasant for the viewer. So it is important to ensure, and especially at low light levels, that the voltage on C2 and therefore the current through the LEDs is as constant as possible. Leading Edge and Trailing edge dimmers as used today, provide just 2 sets of charge to the LEDs over each mains cycle. For Leading and Trailing edge dimmers the charge timing either starts or ends with a zero crossing of the mains cycle. At veiy low levels the Smoothing Capacitor across the LEDs (C2 in Figure 9) provides ample smoothing as very little current is required. At higher output levels the multiple times that power is supplied to the LEDs using the novel technique described in this document compared to the Trailing and Leading Edge type dimmers will produce a smoother current through the LEDs and hence a more constant light output from the LEDs. To fully explain the workings of the Dimmer as exemplified in Figure 9 consider Figure 19 which shows the voltages of the signals at anodes of DI and D3 (SI) and the control signals for Ml over about 2 mains cycles. When SI has reached its lowest value and starts to rise, even though the control signal is off, it shows that SI increases until it reaches about 90V above its minimum value, it then stops increasing until the control signal switches on and then jumps to 270V above the minimum. This behaviour occurs because when D4 and D32 are not conducting, then the impedance seen by Cl is veiy high so there will be no voltage change across Cl, regardless of the state of Ml. Once D4 or D32 start to conduct, then the impedance drops rapidly and the voltage on Cl depends upon the state of Ml and the voltage input from the mains. By dividing the serially connected LEDs into blocks of LEDs, each with fewer LEDs, and driving each block separately with its own Feed Capacitor, Full Wave Rectifier and Smoothing Capacitor, the 90V initial value can be reduced significantly. (See Figure 17 for an example of how this can be achieved). This will allow more switch-ons and switch-offs to be organised, thereby resulting in a less variable voltage on the Smoothing Capacitors. Considering Figure 6, the input Voltage waveform. The thick black line is a reasonable estimate of the shape of a mains voltage waveform with an overlay showing various points to be explained below. In order to provide the required average current to the LEDs, the input from the mains is switched on and off by a suitable device at defined times. An example of a device that can operate as a Switching Unit is a MOSFET or an IGBT These devices are examples of devices that can operate as unipolar switches controlled by a Control unit. This switch is shown as Ml in the Figure 9 is a MOSFET. An IGBT (Insulated Gate Bipolar Transistor) would also do a similar job. These devices can only operate when the Voltage between the Drain and Source or between the collector and emitter is unipolar. That is why a Full Wave Rectifying bridge circuit is used (DI, D2, D3 and D21) in Figure 9. If a suitable bipolar switch would be available then the Full Wave Rectifying bridge circuit would not be necessary. The Switch Ml is controlled by a voltage supplied by V2 which is the controlling unit. This could be any circuit which provides the necessaiy control voltage or current to Ml in Figure 9 at the required time. An example of a suitable unit is a microprocessor which can generate the control signals required. A voltage or current converter may be required between the Control unit V2 and the switch Ml depending on the actual devices used. In the example of when an IGBT or MOSFET is used, the control signal at the gate needs to be relative to the voltage at the emitter or drain and that is why the other terminal of the control unit V2 is shown connected to the drain of Ml in Figure 9. Voltages are then relative to the drain of the MOSFET or the emitter of an IGBT. In order that the output of the Control device V2 is switched on and off at the correct time relative to the Input waveform of the Mains, it needs to be synchronised to the Mains input. When it is synchronised to the incoming Mains signal it will then be able to provide control signals to Ml or XI so that it will be switched on and off at the required time relative to the incoming Mains signal. This synchronisation needs to be accurate so that the average current through the LEDs remains smooth and the light output of the LEDs appears constant to the human eye. An example of a circuit which will provide a suitable signal which can be used by the controller V2 is shown in Figure 13. An engineer versed in the art of electronics would be able to modify this circuit to suit particular needs, and this circuit is given as an example of a circuit which can operate satisfactorily in a domestic environment with low noise present on the Mains signal. In Figure 13 the input is the 230VAC Mains input used and the output is a rectangular wave which has a High value at the Low Voltage DC supply and a low level at GND, which is directly time related to the phase of the incoming Mains signal. If this output signal is used as an input to a microcontroller, then the microcontroller can then be used to generate control signals for the Ml or XI switch which are directly synchronised to the incoming Mains signal. Figures 7 and 8 show examples of how the Switching on and switching off could be positioned in the Mains AC cycle. In Figure 7 the Switching-on and switching-off take place about equally before and after the zero crossing at times B and A respectively. While in Figure 8, the switching on and off at the positions shown by A, B, D and E are the same on the positive part of the cycle and the negative part of the cycle but C and F are not the same. These switch-on and switch-off sequences will produce voltage pulses through the Full Wave Rectifier Bridge comprised of D4, D22, D20 and D23 in Figure 9 and let the average charge produced by these pulses over one mains cycle across C2 can be denoted by QdimB. While the switch Ml is on (Low impedance), the charge supplied across C2 in Figure 9 can be equated to the charge supplied to Cl when the mains supply is directly connected to C2 in the nondimming case Figure 5 over the time the Switch is on. When the switch Ml is switched off (High impedance), this supply is no longer there, so the average charge supplied across C2 is less than in the non-dimming case. Let this average charge over a mains cycle is Qdimi. Taking the two sources of the charge together, we obtain a figure for the average charge supplied to the C2 as being Qtotai where: Qtotal= Qdimi +QdimB Since it is necessary to switch-off and switch-on the switch Ml in order to control the average charge being supplied across C2, we cannot avoid the effects of that switch-on. A very rapid switch on and will sometimes produce RF emissions which need to be controlled. Of course the slope of the switching on and off can be modified, but that will lead to a loss in efficiency and increase in the thermal losses in the switching process and we wish to avoid this source of inefficiency as much as possible. It should be noted that this inefficiency will be in the Dimmer which is usually not included in the Lamp but positioned at a distance from the lamp, so any heat produced by this inefficiency will be produced in the Dimmer and not the Lamp. A simple RC circuit at the output of V2 provides an example of a method which can be used to effect this reduction in pulse slope (See Figure 9 RI and C3). The values of QdimB depend upon several factors. These include pulse slope, the voltage change at the anode of D3 when the switch changes from its OFF state to its ON state, and the value of Cl in Figure 9. A change in any of these factors produces changes in the value of QdimB. In order to provide a steady light output form the LEDs, it is important to keep QdimB and Qdimi non-varying. Good synchronisation between the AC waveform and the switch-on and switch-off of the controller is important to reduce variability and the circuit shown in Figure 13 is an example of a suitable circuit. More charge Q can be added to C2 by switching on and off the switch Ml or XI at other parts of the mains cycle. Filling in between existing Switch-ons and switchoffs will reduce the variation in voltage across C2 in Figure 9 and reduce the peak to average range of the current through the LEDs, because at each switch-on a charge is added to the Capacitor C2. The amount of charge added depends upon the rate of switch-on and when the switch-on takes place during the AC cycle. In this case it may be then possible to reduce the value of C2 while still keeping the same variation of voltage across C2. The circuit designer can therefore choose between a larger value for C2 or a greater variation in the value of voltage across C2. This method of reducing the variation of voltage across C2 may be used to reduce the requirement for an electrolytic capacitor as C2. Often electrolytic capacitors would be chosen for C2, but they do have a limited lifetime, and the elimination of the need for an electrolytic capacitor will usually increase the lifetime of the circuitry. The use of extra switch-ons and switch-offs such as those shown in Figure 8 in positions marked as A and D would only be needed when the current flow in the LEDs is more than at a low level. Therefore the number of switch on and off sequences per mains cycle can vaiy and be dependent on the current requirements for the LED. The number of switch-ons and switch-off which can be used also depends on the voltage required across the LEDs, so a lower voltage across the LEDs will permit more switch-ons and switch-off, resulting in a smaller peak to average voltage across the cycle and a smoother light output. This can be achieved by dividing up the required number of LEDs into “Blocks”. If the total number of LEDs required is divided up into Blocks , each Block would need its own Full Wave Rectifier Bridge, own Feed Capacitor and its own Smoothing Capacitor. This configuration is shown for 3 blocks of LEDs in Figure 17. At veiy low level probably only 2 sequences per mains cycle would be used, but at high level, the number of switch-on and switch-off cycles may be increased so that the peak to average voltage across C2 remains is reduced. The positions of these switch-ons and switch-offs in the Mains cycle will depend upon the design requirements as well as the power output required. Changing the number of switch-ons and switch-offs through one mains cycle is a novel technology. A further novel feature is to change the position of the switch-on and switch-off to adjust the current through the LED to meet the requirements of the lighting needs. As examples of the possibilities of using this technique of switching-on and switching off at various parts of the Mains cycle, consider Figures 14, 15 and 16. In Figure 14, switch on takes place 1ms before the zero crossing and switch off takes place 1ms after the zero crossing. This switching produces a RMS current through the LEDs of 11.5mA. Bringing forward the switching on to 2ms before the zero crossing and delaying the switching off to 2ms after the zero crossing produces 23.5mA through the LEDs (Figure 15). Adding an extra switch on 0.5ms before the peak and switching off 0.5 after the peak produces a total of 38mA through the LEDs (Figure 16). It should be noted that providing the extra current at the peaks has reduced the peak to average value and will provide for a smoother light output from the LEDs. To minimise the peak to average current through the LEDs, it is important to spread the charge distribution equally over the whole AC cycle. As discussed earlier, the temperature of electronic components have a significant affect on their operating life and the components on a lamp should be protected against high temperatures. If a component with a temperature sensitive parameter (Internal Temperature Sensitive Component in Figure 4) is placed in the lamp with the LEDs, it is possible to determine the ambient temperature by measuring the parameter. If this parameter is measured by the Internal Control unit in Figure 4, and a conversion table is available to relate the parameter to temperature then, by measuring the parameter the Internal Control unit can determine then if or when the ambient temperature is deemed to have surpassed a previously determined value, then the charge being sent to the LEDs can be reduced by adjusting the switch-ons and switch-off sequences so that the heat produced in the lamp is reduced and the temperature drops to below the predetermined value. In the general case, when more than one lamp is attached to the dimmer (as shown in Figure 12 with 3 lamps), the current in each is slightly reduced due to the losses in the switch Ml (See Figure 12) when compared to the current through the LEDs when only one lamp is used (Figure 10). Improvements to LED Dimmer and Driver design

Claims

1) A Light Emitting Diode (LED) Dimmer comprising of a switch mechanism which enables the switching-on and switching-off of a Low impedance path from one port of an AC supply to a serial Feed Capacitor, the other end of which is connected to an AC port of a Full Wave Rectifier, the DC output of which is connected to a serially connected plurality of LEDs across which is connected a Smoothing Capacitor, in which the switching-on and switching-off of the Low impedance path are synchronised to the phase of the AC supply, and neither switching function take place at the zero crossing of the AC supply, the other port of the AC supply being connected to the remaining AC port of the Full Wave Rectifier, the switch having the characteristic that it provides a Low impedance to electric current when switched on and a High impedance to electric current when switched off, with the timing of the control of the Switch being synchronised to the AC cycle of the Mains AC supply, in which the switching on or off does not take place at the zero crossing of the Mains AC cycle.2) A LED Dimmer which comprises a first Full Wave Rectifier with a Switch controlled by a controller is connected by one of its AC ports to one connection of a Mains AC supply and the DC output ports are connected together by the Switch, with the other AC port of first Full Wave Rectifier is connected to one or more second Full Wave Rectifiers AC ports via one or more serial Feed Capacitors, with the other AC port of the second Full Wave Rectifier is directly connected to the other connection of the Mains AC input, the DC ports of the second Full Wave rectifier being connected to a serially connected plurality of Light Emitting Diodes (LEDs), the Switch being controlled by the controller which can switch on and switch off the Switch at any time during the AC cycle, the Switch having the characteristic that it provides a Low impedance to electric current when switched on and a High impedance to electric current when switched off, with the timing of the control of the Switch being synchronised to the AC cycle of the Mains AC supply, in which the switching on or off does not take place at the zero crossing of the Mains AC cycle.3) A LED Dimmer which comprises a first Full Wave Rectifier with a Switch controlled by a controller is connected by one of its AC ports to one connection of a Mains AC supply and the DC output ports are connected together by the Switch, with the other AC port of first Full Wave Rectifier is connected to one or more second Full Wave Rectifiers AC ports via one or more serial Feed Capacitors, with the other AC port of the second Full Wave Rectifier is directly connected to the other connection of the Mains AC input, the DC ports of the second Full Wave rectifier being connected to a serially connected plurality of Light Emitting Diodes (LEDs), the Switch being controlled by the controller which can switch on and switch off the Switch at any time during the AC cycle, the Switch having the characteristic that it provides a Low impedance to electric current when switched on and a High impedance to electric current when switched off, with the timing of the control of the Switch being synchronised to the AC cycle of the Mains AC supply, in which the switching on or off takes place at least three times in each Mains AC cycle.4) A Dimmer as described in Claim 1, Claim 2 or in Claim 3 in which Smoothing Capacitors are connected in parallel with the plurality of LEDs.5) A Dimmer as described in Claim 1 or in Claim 2 or in Claim 3 or in Claim 4 which is implemented with a single LED.6) A Dimmer arranged as in Claims 1, 2, 3 or 4 when the number and position of the switch-ons and switch-offs is produced in response to the need to adjust the average current through the LEDs.7) A Dimmer arranged as in Claims 1, 2, 3 or 4 when the additional switch-ons and switch-offs is produced in response to the need to reduce the peak to average voltage across the LEDs.8) A Dimmer arranged as in any of the previous claims so that a temperature sensitive device positioned close to the LEDs can be used to modify the switch-ons and switch-offs to reduce the current through the LEDs when the temperature is deemed to be above a pre-determined temperature.9) A Dimmer arranged as in any of the previous claims when the additional switch-ons and switch-offs are arranged throughout the AC cycle so that the charge sent to the LEDs is approximately equal and equally spaced throughout the cycle.10) A Dimmer arranged as in any of the previous claims when the LEDs in each Lamp are divided into blocks with each block being connected to the First Full Wave Rectifier with its own Feed Capacitor and its own Second Full Wave Rectifier.A

Images