A New Design Method for DC Drive Circuit of LED Lighting

fighting

A New Design Method for DC Drive Circuit of LED Lighting [Copy link]

A New Design Method for DC Drive Circuit of LED Lighting

This article describes a new method for driving LEDs in DC lighting systems that provides 95% efficiency, longer life, and greater resistance to electrical and mechanical shock. A practical circuit design using the ZXSC300 series DC-DC controller is calculated and analyzed.

LED lighting is becoming increasingly popular as an alternative technology to halogen low-voltage lighting. Unlike halogen bulbs, LEDs do not suffer from low efficiency, poor reliability, and short life. This article describes a new method for driving high-power LEDs in DC lighting systems. This solution provides 95% efficiency, longer life, and the ability to withstand higher electrical and mechanical shocks.

Figure 1: LED driving using a buck-mode DC-DC converter.

In the circuit shown in Figure 1, the ZXSC300 series DC-DC controller drives an external switch that works in buck mode. Table 1 lists the bill of materials for a 12V power system. A higher system voltage can be provided by increasing the value of R2. For example, to obtain a 24V voltage, only the value of R2 needs to be changed to 2.2kΩ. At the same time, capacitor C1 must also have a higher rated voltage. The basic working principle of the circuit is as follows:

When Q1 is turned on, current flows through the LED, capacitor C2, and inductor. When the voltage drop across R1 reaches the threshold voltage of the I _sense pin, Q1 turns off for a fixed time, and the energy in the inductor flows through D1 and the LED. After this fixed time, Q1 turns on again, and the cycle repeats.

Circuit working principle analysis

The working principle of the circuit is analyzed in more detail to obtain the circuit parameters and calculations related to system design. The analysis starts with the switch Q1 being turned on for a fixed time T _ON . The ZXSC310 turns Q1 on until it detects a voltage of 19mV (nominal value) on the I _sense pin, so the current on Q1 when this threshold voltage is reached is 19mV/R1, called I _PEAK .

When Q1 is turned on, current flows from the power supply through C1 and the series LED. Assuming the forward voltage drop of the LED is V _F , the remaining power supply voltage will all fall on L1, called V _L1 , and cause the current on L1 to rise at a slope of di/dt=V _L1 /L1 . Where di/dt is in ampere/second, V _L1 is in volts, and L1 is in henry.

The voltage drop across Q1 and R1 is negligible because the on-resistance RDS(ON) of Q1 is very small and the voltage drop across R1 is always less than 19mV. 19mV is the turn-off threshold voltage of Q1, which is set according to the threshold voltage of the I _{sense pin.}

V _IN =V _F +V _L1 T _ON =I _PEAK xL1/ V _L1

Since the voltage across L1 is obtained by subtracting the LED forward voltage drop from V _{IN , T}_ON can be calculated. Therefore, if L1 is smaller, T _ON is also smaller for the same peak current I _PEAK and supply voltage V _IN . Note that during the process of the inductor current rising to I _PEAK , current flows through the LED, so the average current in the LED is equal to the sum of the current during the T _ON rising period and the T _OFF falling period.

Now let's look at the off period of Q1 (T _{OFF ). The T}_OFF of the ZXSC300 series DC-DC controller is internally fixed to 1.7us (nominal value). It should be noted that if this value is used to calculate the current ramp, its range is a minimum of 1.2μs and a maximum of 3.2μs.

To minimize conduction loss and switching loss, T _ON cannot be much smaller than T _OFF . Too high a switching frequency will result in higher dv/dt, so the recommended maximum operating frequency for ZXSC300 and 310 is 200 kHz. Assuming a fixed T _OFF of 1.7μs, the minimum T _ON is 5μs-1.7μs=3.3μs. However, this is not an absolute limit, and these devices can already operate at 2 to 3 times this frequency, but the conversion efficiency will be reduced.

Figure 2: Typical performance curves for a 12V system.

To minimize conduction loss and switching loss, T _ON cannot be much smaller than T _OFF . Too high a switching frequency will result in higher dv/dt, so the recommended maximum operating frequency for ZXSC300 and 310 is 200 kHz. Assuming a fixed T _OFF of 1.7μs, the minimum T _ON is 5μs-1.7μs=3.3μs. However, this is not an absolute limit, and these devices can already operate at 2 to 3 times this frequency, but the conversion efficiency will be reduced.

During T _OFF , the energy stored in the inductor will be transferred to the LED, with only some losses in the Schottky diode. The energy stored in the inductor is:

EQ1

The system can operate in either continuous or discontinuous mode, the difference between the two and the effect on the average current will be explained in the following section.

If TOFF _is exactly the time it takes for the current to reach zero, the average current in the LED will be IPEAK _/ 2. In reality, the current may reach zero before TOFF _, in which case the average current will be less than _IPEAK / 2 because there is a period of time during this cycle when the LED current is zero. This is called "discontinuous" operation.

If the current does not reach zero after 1.7μs, but instead falls to IMIN _, the device is said to be in "continuous" operation. The LED current will rise and fall between IMIN _and IPEAK ₍ with possibly different di/dt slopes), with the average LED current being the average of _IMIN and _IPEAK .

The above principle can be applied to actual circuit design by calculating with actual values. For example, if a 12V DC power supply with stable output voltage and three 1W LEDs (requiring 340mA operating current) are known, the circuit shown in Figure 1 and the bill of materials listed in Table 1 can be used for design. This design can operate within the power supply voltage range of 11V to 18V.

Power input voltage = V _IN = 12V, LED forward voltage drop = V _F = 9.6V, V _IN = V _F + V _L1 . Therefore, V _L1 = 12V-9.6V=2.4V.

Peak current = V _sense / R1 = 34mV/50m(= 680mA, where R1 is R _sense .

T _ON =I _PEAK x L1/V _L1

In the above equations, the LED forward voltage drop is approximated to remain constant throughout the current rise and fall. In fact, it will increase with increasing current, but these formulas make the design calculation results within the tolerance range of the components used in the actual circuit. In addition, the difference between V _IN and V _F is less than either of them, so the 6.2μs rise time will basically depend on these voltage values.

It is worth noting that for a 9.6V LED forward voltage drop and a 300mV Schottky diode forward voltage drop, the time to fall from 680mA to zero is:

Since TOFF _is typically 1.7μs, the current has plenty of time to fall to zero. However, although 1.5μs is pretty close to 1.7μs, the coil current may not fall to zero due to component tolerances. This is not a big deal, as the residual current will be small. It is important to note that due to the measurement and shutdown of the peak current, the dangerous "inductor staircasing" problem that occurs in converters with fixed TON times is not possible _. Since the current may never exceed IPEAK _, even if the current increases from a finite value (i.e., continuous mode), it will never exceed IPEAK _, so the LED current will be approximately equal to the average of 680mA and 0, or 340mA. It is not strictly an average, because there is a 200ns period when the current is zero, but this is very small compared to IPEAK _and component tolerances.

Figure 2 and Figure 3 describe the performance of 12V and 24V systems respectively.

Circuit design calculation

During T _ON (assuming discontinuous operation), the power supply's input power is equal to V _IN ×I _PEAK /2, so the power supply's average input current is equal to this current multiplied by the ratio of T _ON to the total cycle time.

The above equation shows how the average supply current increases at lower voltages as T _ON increases relative to the fixed 1.7μs. This is consistent with the power principle, because when the supply voltage is lower, a fixed (or nearly fixed) LED power requires more supply current to achieve the same power.

The energy stored in the inductor is equal to the energy transferred from the inductor to the LED (assuming discontinuous operation) and is:

EQ1

Therefore, when the difference between the input voltage and the output voltage becomes larger, the energy transferred from the inductor to the LED is greater than the energy directly obtained by the LED from the power supply. If the inductor value L1 and the peak current I _PEAK can be calculated so that the current reaches zero at exactly 1.7μs , the power of the LED will not be too dependent on the power supply voltage, because the average current in the LED is always approximately I _PEAK /2.

As the supply voltage increases, the T _ON required to reach I _PEAK decreases, but the LED power is essentially constant, and only draws supply current from zero to I _{PEAK during T}_{ON . The higher the supply voltage, the smaller the proportion of T}_ON to the entire cycle, so the average supply current is also smaller at higher supply voltages, thus keeping the power (and efficiency) constant.

The forward voltage drop of the Schottky diode reduces efficiency. For example, if the V _F of the LED is 6V and the V _F of the Schottky diode is 0.3V, the efficiency loss of the energy transferred from the inductor is 5%, which is the ratio of the forward voltage drop of the Schottky diode to the forward voltage drop of the LED. During T _ON , the Schottky diode is not in the current loop and does not introduce losses, so the overall efficiency loss ratio depends on the ratio of T _ON to T _{OFF . For low supply voltages where T}_ON accounts for most of the entire cycle, the losses introduced by the Schottky diode are not large. When the LED voltage is higher (multiple LEDs in series), the losses introduced by the Schottky diode are also not large because the forward voltage drop of the Schottky diode will account for a smaller proportion of the total voltage drop.

Conclusion

The circuit design in this article shows how to use a high-efficiency circuit to drive LEDs in halogen bulb replacement applications. Although LEDs have a higher initial cost than halogen bulbs, the total cost is lower or equivalent. In some applications where replacement is difficult or expensive, LEDs may be the only cost-effective solution. As LED lighting output efficiency gradually increases and costs decrease, the trend toward using LED lighting will become more obvious.

By: Ho Wong
Product Marketing Manager
Zetex (Asia) Ltd.

Related link: http://www.eetchina.com/ART_8800363737_865371_cdb25238.HTM

[ Print ] [ Return ]

About the Station Statement Website Introduction Website Promotion Feedback