CONTROL METHOD AND DEVICE EMPLOYING PRIMARY SIDE REGULATION IN A QUASI-RESONANT AC/DC FLYBACK CONVERTER

[0001] 1. Technical Field

[0002] The present disclosure relates to converters and, more particularly, to a control device for quasi-resonant AC/DC flyback converters.

[0003] 2. Description of the Related Art

[0004] Converters, and particularly offline drivers of LED-based lamps for bulb replacement, are often desired to have a power factor greater than 0.9, low total harmonic distortion (THD) and safety isolation. At the same time, for cost reasons, it is desirable to regulate the output DC current required for proper LED driving without closing a feedback loop. In addition, compatibility with dimmers is becoming more and more important for LED drivers, especially dimmers based on phase-cut technology.

[0005] High-power-factor (high-PF) flyback converters are able to meet power factor and isolation specifications with a simple and inexpensive power stage. In a high-PF flyback converter there is not an energy reservoir capacitor directly connected to the input rectifier bridge, so that the voltage applied to the power stage is a rectified sinusoid. To achieve high-PF, the input current tracks the input voltage, thus originating a time-dependent input-to-output power flow. As a result, the output current contains a large AC component at twice the main line's frequency.

[0006] A quasi-resonant flyback converter has the power switch turn-on synchronized to the instant the transformer demagnetizes (i.e. the secondary current has become zero), normally after an appropriate delay. This allows the turn-on to occur on the valley of the drain voltage ringing that follows the demagnetization, often termed “valley-switching.” Most commonly, peak current mode control is used, so the turn-off of the power switch is determined by the current sense signal reaching the value programmed by the control loop that regulates the output voltage or current.

[0007] In a flyback converter the input current is the average of the primary current, which flows only during the ON-time of the power switch, resulting in a series of triangles separated by voids corresponding to the OFF-time of the power switch. This “chopping” causes the average value of the primary current to be lower than half the peak value and depend on the mark-space ratio of the triangles. As a result, the input current is no longer proportional to the envelope of the peaks and unlike the envelope, which is sinusoidal, the input current is not sinusoidal. Although a sinusoidal-like shape is maintained, the input current is distorted. This distorted sinusoidal input current results in a flyback converter that fails to achieve low THD or unity power factor.

[0008] FIG. 1 shows a high-power-factor (high-PF) flyback converter 30 according to the prior art. The hi-PF flyback converter 30 is powered from an AC power line having voltage V_ac(θ) and includes an input bridge rectifier 34 having inputs 32 that receive the voltage V_ac(θ), a first output connected to ground, and a second output at which the rectifier is configured to produce a rectified sinusoidal voltage V_in(θ)=V_PK| sin θ| and the current drawn from the power line is sinusoidal-like.

[0009] On the primary side, the flyback converter 30 also comprises a capacitor C_in, which serves as a high-frequency smoothing filter, connected across the output terminals of the bridge rectifier 34, with the negative end connected to ground, and a voltage divider Ra−Rb. The flyback converter 30 has a transformer 36 with a primary winding L_p, connected to the positive terminal of the capacitor C_in, and an auxiliary winding L_aux coupled to a resistor R_ZCD. A power switch M has its drain terminal connected to the primary winding L_p and its source terminal connected to ground via a sense resistor Rs. The current flowing through the power switch M (i.e. the current flowing through the primary winding L_p when M is ON) can be read as a positive voltage drop across the sense resistor Rs. The primary side of the converter also includes a clamp circuit 37 that clamps leakage inductance of the primary winding L_p.

[0010] On the secondary side, the transformer 36 includes a secondary winding L_s, that has one end connected to a secondary ground and the other end connected to the anode of a diode D having a cathode connected to the positive plate of a capacitor C_out that has its negative plate connected to the secondary ground.

[0011] This flyback converter 30 generates a DC voltage V_out at its output terminals across the capacitor C_out that will supply a load 40, which is a string of high-brightness LEDs in FIG. 1.

[0012] The flyback converter has a divider block 42 having a first input that receives a signal B(θ), and a second input that receives a signal A(θ) that is a portion of the instantaneous rectified line voltage sensed across the capacitor C_in and brought to pin MULT through the resistor divider Ra−Rb. The divider ratio Rb/(Ra+Rb) will be denoted with K_p.

[0013] The capacitor C_T is assumed to be large enough so that the AC component (at twice the line frequency f_L) of the signal B(θ) is negligible, at least to a first approximation, with respect to its DC component B₀.

[0014] The output of the divider block 42 is the division of a rectified sinusoid times a DC level, then still a rectified sinusoid whose amplitude depends on the rms line voltage and the amplitude of the control voltage B₀; this will be a reference voltage Vcs_REF(θ) for the peak primary current.

[0015] The signal Vcs_REF(θ) is fed to the inverting input of a pulse width modulation comparator 44 that receives at its non-inverting input the voltage Vcs(t, θ), sensed across the sense resistor Rs. The voltage Vcs(t, θ) is proportional to the instantaneous current I_p(t, θ) flowing through the primary winding L_p and the power switch M when the switch M is ON. Assuming the power switch M is initially ON, the current through the primary winding L_p will be ramping up and so will the voltage across the sense resistor Rs. When Vcs(t,θ) equals Vcs_REF(θ), the PWM comparator 44 resets the SR flip-flop 46 which switches off the power switch M. Therefore, the output of the divider 42, shaped as a rectified sinusoid, determines the peak value of the current of the primary winding L_p. As a result, the peak value of the primary winding current will be enveloped by a rectified sinusoid.

[0016] After the power switch M has been switched off, the energy stored in the primary winding L_p is transferred by magnetic coupling to the secondary winding L_s and then dumped into the output capacitor C_out and the load 40 until the secondary winding L_s is completely demagnetized. When the secondary winding L_s is demagnetized, the diode D opens and the drain node becomes floating, which was fixed at V_in(θ)+V_R while the secondary winding L_s and the diode D were conducting, with V_R being the reflected voltage seen across the primary winding. The voltage at the drain node would tend to eventually reach the instantaneous line voltage V_in(θ) through a damped ringing due to its parasitic capacitance that starts resonating with the primary winding L_p. The quick drain voltage fall that follows the demagnetization of the transformer 36 is coupled to the pin ZCD of the controller through the auxiliary winding L_aux and the resistor R_ZCD. A zero crossing detector (ZCD) block 48 releases a pulse every time it detects a falling edge going below a threshold and this pulse sets the SR flip flop 46 and drives ON the power switch M, starting a new switching cycle.

[0017] An OR gate 50 between the ZCD block 48 and the set input of the SR flip flop 46 allows the output of a STARTER block 52 to initiate a switching cycle. The STARTER block outputs a signal at power-on when no signal is available on the input of the ZCD block 48 and prevents the converter from getting stuck in case the signal on the input of the ZCD block 48 is lost for any reason.

[0018] The ZCD block 48 also generates a FW signal that is high during transformer's demagnetization, as shown in FIG. 2, and is used by the control loop 56 to generate the B(θ) signal.

[0019] Assuming θε(0, π), according to the control scheme under consideration the peak envelope of the primary current is given by:

I_pkp(θ)=I_p(T_ON,θ)=IPKp sin θ  (1)

[0020] It is worth noticing that this scheme results in a constant ON-time T_ON of the power switch M:

$T_{\mathrm{ON}}=\mathrm{Lp}\frac{I_{\mathrm{PKp}}\sin \theta }{V_{\mathrm{PK}}\sin \theta }=\mathrm{Lp}\frac{I_{\mathrm{PKp}}}{V_{\mathrm{PK}}}$

[0021] For simplicity, the OFF-time T_OFF(θ) of the power switch M will be considered coincident with the time T_FW(θ) during which current circulates on the secondary side. In other words, the time interval T_R during which the voltage across the power switch M rings (starting just after T_FW(θ), as the current in the secondary winding L_s has gone to zero), until reaching the valley of the ringing will be neglected. This is acceptable as long as T_R<<T_OFF(θ).

[0022] The switching period T(θ) is therefore given by:

T(θ)=T_ON+T_FW(θ)

[0023] Considering volt-second balance across the primary winding L_p it is possible to write:

$T_{\mathrm{FW}}\left(\theta \right)=T_{\mathrm{ON}}\frac{V_{\mathrm{PK}}\sin \theta }{V_R}$

[0024] where V_R is the reflected voltage, i.e. the output voltage V_out times the primary-to-secondary turns ratio n=N_p/N_s, seen across the primary winding L_p of the transformer 36 in the time interval T_FW(θ):

V_R=n(V_out+V_F)

[0025] where V_F is the forward drop on the secondary diode D. Therefore:

T(θ)=T_ON(1+K_v sin θ)

[0026] with K_v=V_PK/V_R.

[0027] The input current I_in to the converter 30 is found by averaging the current I_p(t,θ) in the primary winding L_p over a switching cycle. The current I_p(t,θ) is the series of gray triangles in the right-hand side of FIG. 2 so it is found that:

$I_{\mathrm{in}}\left(\theta \right)=\frac{1}{2}I_{\mathrm{pkp}}\left(\theta \right)\frac{T_{\mathrm{ON}}}{T\left(\theta \right)}=\frac{1}{2}I_{\mathrm{PKp}}\frac{\sin \theta }{1+K_v\sin \theta }$

[0028] This equation shows that the input current I_in is not a pure sinusoid: this current is sinusoidal only for K_v=0; when K_v≠0, although a sinusoidal-like shape is maintained, the input current is distorted, the higher K_v the higher the distortion. Since K_v cannot be zero (which would require the reflected voltage to tend to infinity), the prior art QR control scheme does not permit zero Total Harmonic Distortion (THD) of the input current nor unity power factor in a flyback converter even in the ideal case.

[0029] FIG. 3 shows the plots of the THD of the input current and of the power factor versus K_v.

[0030] The regulated DC output current value obtained with this control method is:

$I_{\mathrm{out}}=\frac{{\mathrm{nK}}_D}{2{\mathrm{RsG}}_MR_T}$

[0031] where K_D is the gain of the divider block 42 and G_M the transconductance of a current generator 54 which produces current I_CH(θ).

[0032] This equation shows that with the control method of FIG. 1, which uses only quantities available on its primary side, the DC output current I_out depends only on external, user-selectable parameters (n, Rs) and on internally fixed parameters (G_M, R_T, K_D) and does not depend on the output voltage V_out, nor on the rms input voltage V_in or the switching frequency f_sw(θ)=1/T(θ).

[0033] This control method makes the flyback converter 30 work as a current source. Therefore, even with a chopped AC input voltage—which happens in case the converter is operated through a phase-cut wall dimmer (e.g. leading and trailing edge dimmer as shown in FIG. 5)—the converter forces the preset DC output current to the load.

[0034] In that case, however it would be desirable to reduce the regulation setpoint depending on the dimmer firing angle (1−α) to be compatible with a dimmer: the higher α is, the lower the current set-point should be. This can be realized by modifying the circuit 56 in FIG. 1 as shown in FIG. 4. The sensed input voltage is compared to a threshold voltage V_th in a dimmer comparator 60 and, if it stays below the threshold for a time longer than T_ML, it is assumed that the line voltage is missing (because the dimmer is open) and an EN signal goes low. This freezes the state of the power switch M and disconnects both the current generator 54 producing current I_CH(θ) and the discharge resistor R_T. In this way the voltage across C_T is frozen at the value in the instant when the input voltage goes to zero.

[0035] The delay T_ML prevents the circuit from being improperly activated near the zero-crossings of the line voltage when this is not chopped. Note also that this delay is unidirectional: as the sensed voltage exceeds the threshold voltage V_th the enable signal EN goes high immediately.

[0036] The net effect of stopping the charge/discharge activity of the capacitor C_T can be regarded as an average increase of the discharge resistor R_T, leading to a reduction of the preset output current I_out inversely proportion to the firing angle of the dimmer:

$I_{\mathrm{out}}=\frac{{\mathrm{nK}}_D}{2{\mathrm{RsG}}_MR_T}\left(1-\alpha \right).$

[0037] Real world dimmers have typically a fire angle between 10-20% and 80-90%, and therefore if using the control scheme shown in FIG. 4, the minimum/maximum output current setpoint could be in the range of 10-20% and 80-90% respectively. In other words the control method shown in FIG. 4 cannot meet the typical desired characteristic of a dimmer shown in FIG. 6.

[0038] One embodiment of the present disclosure is a quasi-resonant flyback converter having a sinusoidal input current in order to achieve low total harmonic distortion and high power factor.

[0039] One embodiment of the present disclosure is directed to a control mechanism that enables high power factor (Hi-PF) quasi-resonant (QR) flyback converters with peak current mode control using only quantities available on its primary side able to ideally draw a sinusoidal current from the input source and with an with optimized compatibility to the phase-cut wall dimmers.

[0040] One embodiment of the present invention is a device for controlling a power transistor of a power stage. The device includes a divider having a first input, a second input and an output, the divider being configured to produce a voltage reference signal. A first current generator configured to produce an output current. A shaper circuit configured to output to the first input of the divider a first signal based on the output current of the first current generator. A bias circuit coupled to the first current generator and configured to output a second signal to the second input of the divider; and a driver circuit having a first input configured to receive the reference signal, and an output configured to drive the power transistor.

[0061] FIG. 7 shows a hi-PF QR flyback converter 100 according to one embodiment of the present disclosure. On the primary side, the QR flyback converter 100 includes a controller 102, a bridge rectifier 104 having inputs 106 coupled to an AC power line that supplies an AC voltage V_ac, an input capacitor C_in, a voltage divider R_a−R_b coupled to the bridge rectifier 104, a primary winding L_p and an auxiliary winding L_aux of a transformer 108, power switch M coupled to the transformer 108 and controlled by controller 102, sensing resistor R_s coupled to the power switch M and controller 102, a resistor R_ZCD coupled to the auxiliary winding L_aux, and a clamp circuit 109 connected across the primary winding L_p.

[0062] On the secondary side of the converter 100, a secondary winding L_s of the transformer 108 has one end connected to a secondary ground and the other end connected to the anode of a diode D having the cathode connected to the positive plate of a capacitor C_out that has its negative plate connected to the secondary ground. The converter 100 provides an output voltage V_out that supplies power to a load 110, which in FIG. 7 is a set of LEDs, although other loads could be supplied by the converter 100.

[0063] The controller 102 has a reference voltage circuit 116 that is configured to produce a reference voltage V_CSREF and includes a bias circuit 118 and a shaper circuit 120. The controller 102 also includes a driver circuit 121 having a PWM comparator 122, an SR flip-flop 124, an OR gate 126, and a driver 127 configured to drive the power switch M. The PWM comparator 122 includes an inverting input that receives the reference voltage V_CSREF, a non-inverting input that receives a sense voltage V_CS from the sense resistor R_s, and an output that provide a reset signal to a reset input R of the flip-flop 124. The flip-flop 124 also includes a set input S, coupled to an output of the OR gate 126, and an output that is coupled to an input of the driver 127. The OR gate 126 also has first and second inputs coupled to respective outputs of a starter block 128 and a ZCD block 130. The OR gate 126 provides a set signal to the set input S of the SR flip flop when the ZCD block 130 detects a falling edge go below a threshold, or when the starter block 128 produces a start signal as discussed above.

[0064] The reference voltage circuit 116 has a bias circuit 118 and a shaper circuit 120. The shaper circuit 120 has a first current generator 140, a resistor R_t1 coupled to an output of the first current generator 140, a switch 132 that switchably couples the resistor R_t1 to ground, and a capacitor C_t1 coupled between the output of the current generator 140 and ground. The first current generator 140 has an input coupled to a supply terminal Vcc and a control terminal coupled to the voltage divider R_a−R_b via the pin MULT and produces a current I_CH1(θ). The switch 132 is controlled by the output Q of the flip-flop 124 and thereby connects the capacitor C_t1 in parallel with the switched resistor R_t1 when the power switch M is ON.

[0065] The bias circuit 118 includes a second current generator 142 having an input coupled to the supply terminal Vcc, a control terminal coupled to the output of the first current generator 140, and an output at which the second current generator produces a current I_CH(θ). A second switched resistor R_t is switchably coupled to the output of the second current generator 142 by a switch 134 configured to connect the resistor R_t to the second current generator 142 under the control of the signal FW provided by the ZCD block 130. The signal FW is high when the current is flowing in the secondary winding L_s. Another switch 144 is coupled to the output of the second current generator 142 and is configured to connect the output of the second current generator 142 to ground when the ZCD block 130 under control of a signal FW that is an inverted version of the signal FW.

[0066] The reference voltage circuit 116 also includes a divider block 146 having a first input that receives a signal A(θ) from the shaper circuit 120, a second input that receives a signal B(θ) from the bias circuit 118, and an output at which the divider provides the reference voltage V_CSREF.

[0067] The signal A(θ) is generated by the first current generator 140 acting on the switched resistor R_t1 and capacitor C_t1. The current I_CH1(θ) produced by the current generator 140 is proportional to a rectified input voltage V_in produced at the voltage divider R_a−R_b.

[0068] The resistor R_t1 is connected in parallel to the capacitor C_t1 by the switch 132 when the signal Q of the SR flip flop 124 is high, i.e. during the on-time of the power switch M, and is disconnected when Q is low, i.e. during the off-time of the power switch M. The voltage developed across the capacitor C_t1 is A(θ) and is fed to the first input of the divider block 146.

[0069] The current I_ch1(θ) provided by the current generator 140 can be expressed as:

I_ch1(θ)=g_m1K_p(V_PK sin θ)

[0070] where g_m1 is the current-to-voltage gain of the first current generator 140.

[0071] An assumption is that T(θ)<<R_t1 C_t1<<1/f_L. In this way, the switching frequency ripple across the capacitor C_t1 is negligible and I_ch1(θ) can be considered constant within each switching cycle.

[0072] The A(θ) voltage developed across C_t1 by charge balance is:

$A\left(\theta \right)=R_{t1}I_{\mathrm{ch}1}\left(\theta \right)\frac{T\left(\theta \right)}{T_{\mathrm{ON}}\left(\theta \right)}=R_{t1}g_{m1}K_p\left(V_{\mathrm{PK}}\sin \theta \right)\frac{T\left(\theta \right)}{T_{\mathrm{ON}}\left(\theta \right)}$

[0073] The generation of the other input signal B(θ) to the divider block 146 is similar to that of the B(θ) of FIG. 1. The current I_CH(θ) provided by the second current generator 142 and used to generate the B(θ) signal, can be expressed as:

I_CH(θ)=G_MA(θ)

[0074] where G_M is the current-to-voltage gain of the second current generator 142.

[0075] Now considering the C_T by charge balance, it is possible to find the voltage B(θ) developed across the capacitor C_T:

$B\left(\theta \right)=G_MR_Tg_{m1}R_{t1}K_p\left(V_{\mathrm{PK}}\sin \theta \right)\frac{T_{\mathrm{FW}}\left(\theta \right)}{T_{\mathrm{ON}}\left(\theta \right)}$

[0076] The capacitor C_T is assumed to be large enough so that the AC component (at twice the line frequency f_L) of the signal B(θ) is negligible with respect to its DC component B₀, which can be written as:

$B_0=\stackrel{\_}{B\left(\theta \right)}=\frac{1}{\pi }G_MR_Tg_{m1}R_{t1}K_pV_{\mathrm{PK}}{\int }_0^{\pi }\sin \theta \frac{T_{\mathrm{FW}}\left(\theta \right)}{T_{\mathrm{ON}}\left(\theta \right)}\theta =\frac{G_MR_Tg_{m1}R_{t1}K_pV_{\mathrm{PK}}K_V}{2}$

[0077] Considering the voltage-second balance for transformer 108, the primary on time T_ON(θ) and secondary on time T_FW(θ) can be expressed by the following relationship:

$\frac{T_{\mathrm{FW}}\left(\theta \right)}{T_{\mathrm{ON}}\left(\theta \right)}=K_v\sin \theta$

[0078] The voltage reference Vcs_REF(θ) is therefore:

${\mathrm{Vcs}}_{\mathrm{REF}}\left(\theta \right)=K_D\frac{A\left(\theta \right)}{B\left(\theta \right)}\approx K_D\frac{A\left(\theta \right)}{B_0}=K_D\frac{2}{G_MR_TK_v}\sin \theta \frac{T\left(\theta \right)}{T_{\mathrm{ON}}\left(\theta \right)}$

[0079] where K_D is the gain of the divider block 146 and it is dimensionally a voltage. Considering that the peak primary current I_pkp(θ) can be expressed as:

$I_{\mathrm{pkp}}\left(\theta \right)=\frac{{\mathrm{Vcs}}_{\mathrm{REF}}\left(\theta \right)}{\mathrm{Rs}}$

[0080] The input current can be expressed as:

$I_{\mathrm{IN}}\left(\theta \right)=\frac{1}{2}I_{\mathrm{PKP}}\left(\theta \right)\frac{T_{\mathrm{ON}}\left(\theta \right)}{T\left(\theta \right)}$ $I_{\mathrm{IN}}\left(\theta \right)=\frac{K_D}{G_MR_TK_V}\sin \theta \frac{1}{R_S}$

[0081] This results in a sinusoidal input current in a constant-current primary-controlled Hi-PF QR flyback converter 100.

[0082] Considering that the secondary current is n=Np/Ns times the primary current, the peak secondary current I_pks(θ) can be calculated as:

$I_{\mathrm{pks}}\left(\theta \right)={\mathrm{nK}}_D\frac{2}{G_MR_TK_v}\sin \theta \frac{T\left(\theta \right)}{T_{\mathrm{ON}}\left(\theta \right)}\frac{1}{R_S}.$

[0083] Since the cycle-by-cycle secondary current Is(t,θ) is the series of triangles shown in left-hand side of FIG. 8, its average value in a switching cycle is:

$I_o\left(\theta \right)=\frac{1}{2}I_{\mathrm{pks}}\left(\theta \right)\frac{T_{\mathrm{FW}}\left(\theta \right)}{T\left(\theta \right)}=\frac{{\mathrm{nK}}_D}{G_MR_TK_v}\sin \theta \frac{T_{\mathrm{FW}}\left(\theta \right)}{T_{\mathrm{ON}}\left(\theta \right)}\frac{1}{R_S}.$

[0084] The DC output current I_out is the average of I_o(θ) over a line half-cycle:

$I_{\mathrm{out}}=\stackrel{\_}{I_o\left(\theta \right)}=\frac{1}{\pi }{\int }_0^{\pi }\frac{{\mathrm{nK}}_D}{G_MR_TK_vR_S}\sin \theta \frac{T_{\mathrm{FW}}\left(\theta \right)}{T_{\mathrm{ON}}\left(\theta \right)}\theta .$

[0085] Finally, the average output current is:

$I_{\mathrm{out}}=\frac{{\mathrm{nK}}_D}{2G_MR_TR_S}.$

[0086] The previous expression shows that the circuit of FIG. 7 has a DC output current Iout that depends only on external, user-selectable parameters (n, Rs) and on internally fixed parameters (G_M, R_T, K_D) and does not depend on the output voltage Vout, nor on the RMS input voltage Vin or the switching frequency f_SW(θ)=1/T(θ).

[0087] Therefore, it is possible to conclude that the converter 100 of FIG. 7, in addition to providing ideally unity power factor and zero harmonic distortion of the input current, also provides a regulated Iout using only quantities available on the primary side.

[0088] FIGS. 12 and 13 show simulation results of the signals of FIG. 7 with Vin being 265 VAC and 90 VAC respectively, including A(A), B(θ), Iout, Iin, V_CSREF, and the THD of the circuit. It is worth noticing the very low distortion level of the input current (around 3.3% at V_in, =90 Vac, around 3.8% at V_in, =265 Vac), due to the input EMI filter and the non-idealities considered both in the controller 102 and the bridge rectifier 104, transformer 108 and power switch M.

[0089] FIG. 8 illustrates several of the waveforms of converter 100 of FIG. 7. On the left-hand side are the waveforms on a switching period time scale, on the right-hand side the waveforms on a line cycle time scale.

[0090] In FIG. 14 are shown the simulation results comparison between the prior art converter 30 and the presently disclosed converter 100 in terms of THD (left) and PF (right). FIG. 15 shows the simulation results comparison in terms of output current regulation.

[0091] FIG. 9 is a reference voltage circuit 118′ according to one embodiment of the present disclose and can be employed instead of the reference voltage circuit 118 of FIG. 7 when it is desired to obtain the dimming curve shown in FIG. 6. The reference voltage circuit 118′ includes the switches 134, 144, second current generator 142, resistor R_T, and capacitor C_T of the reference voltage generator 118 of FIG. 7. Unlike the reference voltage generator 118 of FIG. 7, the reference voltage circuit 118′ includes a phase angle detector 150 having a comparator 151, a delay block 152, and an AND gate 153. The comparator 151 has an inverting input that receives a sensed input voltage from a dimmer, a non-inverting input that receives a voltage threshold V_th, and an output at which the comparator produces a signal a based on a comparison of the sensed input voltage with the voltage threshold V_th. The delay block 152 adds a masking time delay T_MASK and the AND gate 153 outputs an α_MASK signal.

[0092] The reference voltage circuit 118′ also includes a dimming circuit 154 that includes a dimming current generator 155, a switch 156, and a gain block (G_DIM) 157. An extra current L_dim is added on the B(θ) signal from dimming current generator 155. This current L_dim is proportional to the signal B(θ) and, as shown in FIG. 10, is added only during a part of the dimmer off-time (basically only when α_MASK signal is high and closes the switch 156).

[0093] The reference voltage circuit 118′ further includes inverters 158, 159, a switch 160, and another AND gate 161. The inverter 158 is connected between an output of the AND gate 153 and a control terminal of the switch 160, and thereby, controls the switch 160 based on an inverted version of the α_MASK signal output by the phase angle detection circuit 150. The inverter 159 is connected between an output of the AND gate 161 and a control terminal of the switch 144. The AND gate 161 has first and second inputs connected respectively to the output of the ZCD block 130 that provides the FW signal and the output of the inverter 158 that provides the inverted version of the α_MASK signal. The output of the AND gate 161 is also connected to a control terminal of the switch 134, so the AND gate 161 opens one of the switches 134, 144 while closing the other one of the switches 134, 144, and vice versa, depending on the FW signal output by the ZCD block 130 and on the inverted version of the α_MASK signal provided by the inverter 158.

[0094] The I_DIM current generator 155 is added on the C_T capacitor, increasing the B(θ) signal in function of the dimmer firing angle, resulting in a lower DC output current. In other words, the I_DIM current generator 155 increases the equivalent R_T discharging resistor based on the dimmer firing angle.

[0095] Considering the C_T charge balance, it is possible to find the equivalent discharging resistor:

$R_{\mathrm{Tequivalent}}=R_T\left[\frac{R_{\mathrm{DIM}}}{R_{\mathrm{DIM}}\left(1-{\alpha }_{\mathrm{MASK}}\right)-R_T{\alpha }_{\mathrm{MASK}}}\right]$

[0096] The DC output current is therefore:

$I_{\mathrm{out}}\left[{\alpha }_{\mathrm{MASK}}\right]=\frac{{\mathrm{nK}}_D}{2R_SG_MR_T}*\left[\frac{R_{\mathrm{DIM}}\left(1-{\alpha }_{\mathrm{MASK}}\right)-R_T{\alpha }_{\mathrm{MASK}}}{R_{\mathrm{DIM}}}\right]$

[0097] where

${\alpha }_{\mathrm{MASK}}=\alpha -\frac{T_{\mathrm{MASK}}}{T}$

and T is the line period.

[0098] The previous expression shows that the DC output current depends on the dimmer firing angle (1−α) with a relationship that has a high slope, and can be programmed through the R_DIM resistor. Because of the T_MASK delay time, the DC output current does not change until the dimmer off-time is higher than T_MASK.

[0099] FIG. 11 shows the dimming circuit 154 of FIG. 9 according to one embodiment. The I_DIM current generator 155 is implemented using a control transistor 162 and a current mirror that includes a diode-connected, bipolar first mirror transistor 163 and a bipolar second mirror transistor 164 having respective bases connected to each other and respective emitters connected to the supply terminal Vcc. The dimming circuit 154 also includes a resistor R_DIM and the switch 156 connected in series with the control transistor 162 and the first mirror transistor 163 between the supply terminal Vcc and ground. The switch 156 is implemented as an NPN bipolar transistor having its collector connected to the resistor R_DIM, its emitter connected to ground, and its base connected to the output of the phase angle detector 150 to receive the α_MASK signal. The gain block 157 is implement using an amplifier 165 having its non-inverting input connected to receive the B(θ) signal, its inverting input connected to a node between the emitter of the control transistor 162 and the resistor R_DIM, and its output connected to the base of the control transistor 162.

[0100] FIG. 16 shows simulation results of the circuit of FIG. 9 implemented in the QR converter of FIG. 7. In FIG. 17 is shown a comparison between the prior art converter 30 and the present disclosure converter 100 modified with the circuit of FIG. 9 in terms of dimming curves (output current versus dimmer firing angle).

[0101] Shown in FIG. 18 is an alternative implementation of a shaper circuit 170, which could be used in place of the shaper circuit 120 of FIG. 7 to generate the A(θ) signal. The shaper circuit 170 of FIG. 18 includes the resistor R_t1, capacitor C_t1, and switch 132 of the shaper circuit 120 of FIG. 7 and also includes the resistive voltage divider R_a−R_b of FIG. 7. The shaper circuit 170 also has a current generator 172 connected between the supply terminal Vcc and the resistor R_t1 and configured to supply a current I_ref1. A multiplier block 174 has a first input connected to a node between the output of the current generator 172 and resistor R_t1 and configured to receive a signal A1(θ), a second input connected to the mid-point of the voltage divider R_a−R_b and configured to receive a signal A2(θ) from the voltage divider R_a−R_b, and an output configured to supply the A(θ) signal. Considering the C_t1 charge-balance, the A1(θ) voltage developed across the capacitor C_t1 is:

$I_{\mathrm{ref}1}T\left(\theta \right)=\frac{A1\left(\theta \right)}{R_{t1}}T_{\mathrm{ON}}\left(\theta \right)$

[0102] where I_ref1 is a constant current produced by the current generator 172.

[0103] Considering that A2(θ)=K_p (V_PK sin θ), the A(θ) signal results:

$A\left(\theta \right)=K_MI_{\mathrm{ref}1}R_{t1}K_PV_{\mathrm{PK}}\sin \theta \frac{T\left(\theta \right)}{T_{\mathrm{ON}}\left(\theta \right)}$

[0104] Where K_M is the gain of the multiplier block 174. Comparing the equation for the A(θ) signal produced by the shaper circuit 120 of FIG. 7 with the above equation for the A(θ) signal produced by the shaper circuit 170 of FIG. 18, the implementation shown in FIG. 18 is equivalent to the implementation shown in FIG. 7 if the multiplier gain, K_M, is:

$K_M=\frac{g_{m1}}{I_{\mathrm{ref}1}}$

[0105] Shown in FIG. 19 is alternative implementation of a bias circuit 180, which could be used in place of the bias circuit 118 of FIG. 7 to generate the B(θ) signal. The bias circuit 180 has an amplifier 182 configured to receive the A(θ) signal and produce a signal A1(θ). The amplifier 182 could be configured to receive the A(θ) signal from the shaper circuit 120 of FIG. 2, the shaper circuit 170 of FIG. 18, or a shaper circuit according to an alternate embodiment in view of the above discussion. Also, the amplifier 182 could be implemented by the controlled current generator 140, which produces the current I_ch1(θ) proportionally to the portion of the input voltage V_in(θ) at the midpoint of the voltage divider R_a−R_b, or an alternate amplifier could be employed. A first switch 184 is coupled between the amplifier 182 and the resistor R_t and a configured to connect the amplifier 182 to the resistor R_t based on the FW signal produced by the ZCD block 130. A second switch 186 is coupled between the first switch 184 and ground, and is configured to connect the resistor R_t to ground based on the inverted signal FW.

[0106] One can determine the B(θ) voltage by considering the following C_T charge-balance:

$\frac{A_1\left(\theta \right)-B\left(\theta \right)}{R_T}T_{\mathrm{FW}}\left(\theta \right)=\frac{B\left(\theta \right)}{R_T}T\left(\theta \right).$

[0107] Considering that A₁(θ)=K A(θ), the B(θ) signal is:

$B\left(\theta \right)=\mathrm{KA}\left(\theta \right)\frac{T_{\mathrm{FW}}\left(\theta \right)}{T\left(\theta \right)}$

[0108] where K is the voltage gain of the amplifier 182.

[0109] In FIG. 20 is shown an alternative embodiment of a controller 188, which could be employed in place of the controller 102 of FIG. 7 to control the power switch M. The controller 188 is identical to the controller 102 of FIG. 7 except that the controller 188 includes a shaper circuit 189 instead of the shaper circuit 120. The shaper circuit 189 is configured to implement a line voltage feed-forward in order to eliminate the dependence of the signal B(θ) on the input voltage Vin. The shaper circuit 189 includes the same switch 132, current generator 140, resistor R_t1, and capacitor C_t1 as in the shaper circuit 120 of FIG. 7. In addition, the shaper circuit 189 includes a feed-forward circuit 190, which is composed of a peak detector 192, a quadratic voltage divider 194, and a multiplier 196. The peak detector 192 detects a voltage peak of the portion of the rectified input voltage received from the midpoint of the voltage divider R_a−R_b and provides an output signal representative of that peak. The quadratic voltage divider 194 receives the output signal from the peak detector 192 and produces a feed-forward signal FF equal to:

$\mathrm{FF}=\frac{1}{{\left(K_PV_{\mathrm{PK}}\right)}^2}.$

[0110] The multiplier 196 multiplies the feed-forward signal FF from the quadratic divider 194 to the signal A(θ) produced at the intermediate node between the current generator 140 and the capacitor C_t1 to produce a signal A1(θ):

$A_1\left(\theta \right)=\frac{g_{m1}R_{t1}}{K_PV_{\mathrm{PK}}}\sin \theta \frac{T\left(\theta \right)}{T_{\mathrm{ON}}\left(\theta \right)}.$

[0111] The current I_CH(θ) provided by the current generator 142, used to generate the B(θ) signal, can be then expressed as:

I_CH(θ)=G_MA1(θ).

[0112] Now considering the C_T charge-balance it is possible to find the voltage B(θ) developed across the capacitor C_T:

$B\left(\theta \right)=G_MR_T\frac{g_{m1}R_{t1}}{K_PV_{\mathrm{PK}}}\sin \theta \frac{T_{\mathrm{FW}}\left(\theta \right)}{T_{\mathrm{ON}}\left(\theta \right)}.$

[0113] Finally the DC component of the signal B(θ) is:

$B_0=\frac{G_MR_Tg_{m1}R_{t1}}{2K_P}\frac{1}{V_R}$

[0114] The various embodiments described above can be combined to provide further embodiments. These and other changes can be made to the embodiments in light of the above-detailed description. In general, in the following claims, the terms used should not be construed to limit the claims to the specific embodiments disclosed in the specification and the claims, but should be construed to include all possible embodiments along with the full scope of equivalents to which such claims are entitled. Accordingly, the claims are not limited by the disclosure.