Fundamentals of DC/DC Converters XC9235/XC9236/XC9237 series
Time：2022.06.22
View：
1. Necessity and use of DC/DC converter to suppress ripple
The composition of the power supply
Whether to use a linear voltage regulator or a DC/DC converter
Are you worried about whether to use a linear voltage regulator or a DC/DC converter when discussing the power supply configuration of your machine? When the operating voltage of the LSI drops and operates at 1.8V or 1.2V, such as when a linear voltage regulator is used for driving from a 5V line or a Li-ion battery, a large amount of heat loss will be generated and energy efficiency cannot be used effectively. It is well known that in this case, if a step-down DC/DC converter is used, the voltage can be converted with high efficiency. However, if a DC/DC converter is used, there will be many problems such as [generated noise], [increased cost due to many external parts], [complicated procedures for setting constants], etc. due to [items required for use]. Opinions that hinder use.
In this survey record, we will understand the basic features and working principles of DC/DC converters by explaining the working conditions of TOREX's DC/DC converters XC9235/XC9236/XC9237. In addition, the XC9235/XC9236/XC9237 series has few external parts and is an integrated DC/DC converter that does not require constant adjustment. and other popular ICs.
2. Features of step-down DC/DC converter IC XC9235/XC9236/XC9237
Recently, there have been many products that simply add coils and capacitors to step-down DC/DC converters with few external components. This is because a high-efficiency power supply can be easily obtained by adding only one coil, which is smaller than that of an LDO voltage regulator, to constitute a component.
XC9235/XC9236/XC9237 series
Torex Semiconductor's XC9235/XC9236/XC9237 series have built-in step-down synchronous rectification DC/DC converters using P-channel MOSFET and N-channel MOSFET output current drivers. The circuit constants required for stable operation are determined by the internal IC, and the external components only need the input and output capacitors and coils listed in the product catalog to operate. It is an integrated DC/DC converter. Can provide up to 600mA of output current. (Refer to Figure 1)
This IC has different control methods with different models. XC9235 works for PWM, XC9236 works for PWM/PFM automatic conversion, and XC9237 not only works for PWM/PFM automatic conversion, but also has an additional manual signal fixed to PWM work function. product. Products with two switching operating frequencies of 1.2MHz and 3MHz are available. You can choose 1.2MHz if you need to pay attention to the energy conversion rate; if you give priority to miniaturization, choose a 3MHz product that is smaller than the coil size. In addition, there are a wealth of functions and features that can be selected for a variety of applications.
High-speed soft-start and CL discharge functions can make the rise and fall of turn-on and turn-off proceed quickly, and can easily set the sequence of the combined power supply.
The internal circuit is composed of reference voltage, error amplifier, oscillator, indicator light waveform, PWM comparator, UVLO, limited current circuit, etc. The fixed output voltage type is fixed by laser trimming to adjust the resistance ratio of the internal resistance R1/R2. The working voltage range is between 1.8V and 6V. In addition, the phase compensation of the amplifier, transmission frequency, soft-start time, etc. are adjusted internally, so it can be used without adjustment.
3. Test and usage with XC9235/XC9236/XC9237
PWM/PFM automatic conversion work
Reduce losses
One of the greatest expected purposes of using a DC/DC converter is to convert energy with high efficiency. The way to improve efficiency is to reduce losses. As shown in Figure 3, major causes of loss include IC current consumption, heat loss due to on-resistance of driver transistors, and loss due to series connection of coils.
PFM control operation and PWM control operation
Here, the PFM control works very effectively corresponding to the number of times the load current controls the switching. In PFM operation, when the load current is small, the number of times of switching per unit time is reduced, and the efficiency is improved by reducing the inactive current such as consumption current and through current. This IC suppresses the current consumption under no load to 15μA or less by PFM operation. As the load current increases, the consumption current of the IC becomes relatively small, and a PWM operation with a smaller ripple voltage is required at this time.
PWM/PFM automatic conversion work
XC9236 has the function of automatically switching PFM work and PWM work according to the load change. This feature can simultaneously improve efficiency at light loads and maintain low ripple at heavy loads, which is beneficial to improve efficiency over the entire load range. (Refer to Figure 4)
From Figure 5, the load transient response characteristics at the transition from PFM operation to PWM operation and the reverse transition can be confirmed.
From the load transient response characteristics when the load fluctuates steeply, it can be recognized that the recovery operation starts only at a voltage difference of 20mV, and no overshoot or ringing occurs during recovery, which makes the phase system have sufficient margin.
Manually change working status
On the basis of PWM/PFM automatic conversion, XC9237 adds the function of forced transition to fixed PWM operation by external signal. Corresponding to the working state of the machine, it can work in a PWM working state with lower ripple, and can control the noise at any time (refer to Figure 6-1, Figure 6-2)
Limited current PFM operation
Features of Limited Current PFM
This IC adopts the current-limited PFM control in the PFM operation in order to easily correspond to the low-voltage operation and low-ripple requirements of LSIs used in small portable instruments. The limited current PFM control is a control method in which the switch is turned off when the current flowing through the coil reaches a certain value (180mA).
In the case of a PFM with a fixed duty ratio, when the difference between the input and output voltages increases, the current flowing through the coil increases due to the switching effect, and the ripple also increases. The limited current PFM has the special feature that when the input and output voltage difference increases, the on-time is reduced to reduce the current flowing through the coil, and the ripple shown in Figure 7 also becomes smaller.
In the transition of LSI's operating voltage from 1.8V to lower 1.5V or 1.2V, there will be more stringent requirements for power supply uniformity, accuracy and low ripple.
Maximum on-time limit
In addition, the maximum on-time at PFM has been determined. When the input voltage is close to the output voltage, and the difference between the input and output voltage is very small, even if the switch is turned into an on state and the current flows through the coil, it is difficult to reach 180mA. Limiting the maximum on-time (twice the frequency) can make the switch temporarily become off state, limiting its discontinuous excess on state.
＜＜ Feature / Ripple Noise and Spike Noise ＞＞
Noise generated by DC/DC converters is roughly classified into ripple noise and spike noise.
Ripple noise
Ripple noise is the noise generated by the ESR (equivalent series resistance) of the capacitor and the DC current when the coil current caused by the switching action and the current energy stored in the coil are released to the load capacitor (CL), depending on the type of capacitor. Huge difference. When using tantalum capacitors and aluminum electrolytic capacitors with many ESR components, the noise increases, and when using low-ESR capacitors such as ceramic capacitors, the noise is small, and a sine wave ripple waveform is formed. (Refer to Figure 8)
Spike noise is high-frequency noise generated at the timing of switching transitions, and is mainly caused when the drive transistor starts or stops abruptly. When it is necessary to observe spike noise, attention must be paid to the measurement method. root
Depending on how the measurement wiring is handled, the measurement may not be accurate. Special attention should be paid to the grounding treatment near the probe. In order to prevent the probe from collecting noise and perform processing such as canceling the probe grounding, it is necessary to directly contact the probe terminal with the lead at the shortest distance. In addition, among common failures, there is a lead in the high frequency filter switch of the tester.
An example of measurement in the ON state. The peak noise of the DC/DC converter is high-frequency noise of 20 to 50 MHz. Of course, it cannot be measured when the high-frequency filter is turned on. (Refer to Figure 9)
Applied to circuits that can change the output voltage
Add another resistor to the FB terminal for selection prepared for this IC, and connect a DA converter or an N-channel open-drain to form a variable output voltage circuit as shown in Figure 10.
This is a method for easily adjusting the voltage according to the operating state of a microcomputer or the like. Set the FB pin voltage (VFB) to the state of 0.8V feedback. The output voltage when using the DAC is determined by VOUT=(VFB-VDAC)(RFB1/RFB2)+VFB. Figure 1 illustrates the relationship between VOUT and VDAC.
A square wave and a sine wave are inserted in Figure 11 to evaluate the followability of the VOUT voltage corresponding to the DAC output
If the output voltage has 2 values, it is convenient to use N-channel open drain. (Refer to Figure 12)
When N-channel open-drain is on, VOUT=0.8V×(RFB1+RFB2)/RFB2
When N-channel open-drain is cut off, VOUT=0.8V×(RFB1+RFB2+RFB3)/(RFB2+RFB3)
4. Realize the technology of suppressing ripple
Ripple suppression during automatic PFM/PFM conversion
The biggest feature of this IC is the low ripple voltage when transitioning from PFM operation to PWM operation. When transitioning from light load to heavy load, the pulse interval of the PFM gradually approaches. When the frequency of PWM operation is finally reached, it transitions to the PWM operation state. However, conventional products such as XC9226 cannot generate distinct pulses before reaching the PWM operating conditions, but generate pulses in groups. (Refer to Figure 13)
When the PFM is working, the switch starts to operate when the load current increases and the output voltage drops. At this time, if the output voltage cannot be recovered by one switching operation, the second switching operation will continue. If the 2nd switching action cannot be recovered, continue to perform the 3rd recovery. Since the switching operation is continued until the output voltage is restored, the coil current is locally superimposed and increased, resulting in the output of a large ripple voltage.
Because the PFM switches intermittently, the coil current cannot be distributed evenly, and even in the state of large local current, it is not conducive to the balanced relationship between the output current and the output voltage.
On the other hand, in the case of PWM operation, since the switching operation is performed in a certain period, it is necessary to distribute the coil current evenly at a certain current.
The PWM/PFM automatic conversion control of the XC9236 series dramatically suppresses this localized pulse group, maintains the evenly dispersed current, and can continuously transition from PFM operation to PWM operation, as shown in Figure 14. low ripple state over the load range.
＜< Feature ／ Selecting the switching frequency by weighing the miniaturization and low power consumption ＞＞
Regarding the switching frequency of the DC/DC converter, it is not necessarily the case that the switching operation is fast. The key is to choose the switching frequency according to the purpose of meeting the needs first.
Since high energy conversion efficiency is expected for DC/DC converters, components such as MOSFETs with low on-resistance are used as transistors for current drive. Because the driving method of the MOSFET is capacitive gate drive, the gate is repeatedly charged and discharged with each switching action. Generally speaking, the lower the resistance of the MOSFET, the greater the gate capacity. Accordingly, as shown in FIG. 15 , the higher the switching frequency, the more times, and the ineffective current consumption and the through current are continuously increased, resulting in lower efficiency. If you pay attention to efficiency, it is more effective to choose a low frequency.
One advantage of choosing a high switching frequency is the miniaturization of components. Increasing the frequency can reduce the inductance value set for the coil used, and as a result, reducing the number of turns of the coil and reducing the coil size is beneficial to selection. The reduction in the number of turns of the winding reduces the series resistance, and the diameter of the winding becomes thicker even for the same size coil. Reducing the series resistance component can improve energy efficiency.
In addition, when selecting coils, it is necessary to pay attention to the overlapping characteristics. For example, when an output current of 200mA is required, the peak current flowing through the coil reaches 300mA. (Fig. 14-④) When the overlapping characteristics of the coils are poor, the inclination of the coil current increases and the peak current increases, which eventually leads to an increase in the ripple. At this time, if the limited current starts to work, it may affect the phase control. Coils with overlapping characteristic margins should be chosen liberally.
references
1. Torex Semiconductor TIP (Technical Information Report) No. 00006 "Methods for Evaluating and Reducing Spike Noise"
