Use pulse width modulation to get a precise output voltage
Source: InternetPublisher:武林萌主 Keywords: PWM MAX4544 Updated: 2024/08/30
In recent years, many single-chip microcomputer manufacturers, such as Atmel, Analog DivICes, Intel, Philips, Dallas, Maxim, etc., have launched new high-speed single-chip microcomputers. Their instruction execution cycle is only 1/3~1/10 of the original, and EEPROM, WDT, A/D converter and D/A converter are integrated in the single-chip microcomputer, which greatly improves the performance of the single-chip microcomputer and facilitates users. However, the output of the D/A converter in many single-chip microcomputers adopts the form of pulse width modulation (PWM). PWM is very suitable for the control of switching power supplies, thyristors and other devices, and can also be used in occasions such as LCD brightness control and audio output that do not require accurate voltage output. Since PWM has no reference voltage, the amplitude of its output pulse is not very constant, which limits the scope of use of PWM. In occasions where precise control voltage output is required, such as precision adjustable voltage sources, motor inverters, etc., PWM cannot be used.
However, using only two inexpensive integrated circuits, the PWM output with non-constant amplitude can be converted into a precise PWM output voltage.
1 Circuit Principle
The circuit principle of using a three-terminal precision reference power supply and an analog switch to obtain a voltage-precision PWM pulse is shown in Figure 1. D1 is a TL431 three-terminal reference voltage integrated circuit, U1 uses a single-pole double-throw analog switch MAX4544; resistors R1, R2, and R3 are determined according to specific needs. Of course, other types of integrated circuits can also be used.
When the PWM pulse is at a high level (logic 1), the COM terminal of U1 is thrown to the normally closed terminal (NC), the adjustment pin of TL431 is connected to the positive voltage pin, and the output voltage value is 2.5 V. When the PWM pulse is at a low level (logic 0), the COM terminal of U1 is thrown to the normally open terminal (NO), and the output voltage of TL431 is sent to the adjustment pin after being divided by R2 and R3. At this time, the output voltage value is equal to [(R2+R3)/R3]×2.5 V. In this example, the output voltage is equal to 8 V. In this way, when the PWM signal is input to the IN pin of U1, the circuit outputs a PWM pulse with a high level of 8 V and a low level of 2.5 V, and its amplitude is 8 V - 2.5 V = 5.5 V. If you need to output a PWM signal with a low level of zero, adding a differential amplifier can solve the problem.
In situations where the precision requirement is not very high, a simpler method can be used. Figure 2 is a circuit diagram for using precision voltage regulator diodes to stabilize and limit PWM pulses. In Figure 2, the PWM signal is amplified by the high-speed operational amplifier U1 to become a ±12 V output voltage. After the current limiting of R1 and the voltage stabilization of D1, a ±6.5 V PWM pulse output is obtained.
2 Error analysis
In Figure 1, as long as the reference power supply is selected appropriately, the error of the reference power supply itself can be completely ignored. In addition, the sources of error are mainly the following:
(1) Error caused by the on-resistance of the analog switch
When the analog switch is turned on, there is a certain on-resistance. When the input current of the TL431 adjustment pin passes through the analog switch, a voltage drop will be formed, resulting in an error. The on-resistance of the MAX4544 is 35 Ω, while the input current of the TL431 adjustment pin is less than 4 μA. As a result, the error of the reference voltage is less than 140 μV, which is 0.000 056 of 2.5 V, equivalent to the accuracy of 14 bits of binary.
(2) Error introduced by switching delay time
The switching delay time will cause the pulse duty cycle to change, which will cause errors in the PWM output pulse. The MAX4544 has an on time of 30 ns and an off time of 25 ns. Calculation shows that when the PWM frequency is 10 kHz, the maximum error caused by this is 0.0003, which is equivalent to 12-bit accuracy. If the PWM frequency is selected to be lower, the impact of the switching delay time will be reduced accordingly. For example, when 1 kHz is selected, the error introduced is 0.000 03, which is equivalent to 15-bit accuracy.
Of the two above, the drift of these parameters over temperature and time is what really affects the accuracy of the output power supply. Since the absolute values of these two parameters themselves are very small, it can be inferred that their drift is even smaller.
From the above analysis, it can be seen that the error introduced by the additional circuit can fully meet the accuracy requirement of PWM.
In the circuit of Figure 2, there are three main reasons for the error:
(1) Error introduced by the dynamic resistance of the Zener diode
The dynamic resistance of the Zener diode is relatively large, generally around tens of Ω (when the operating current is 5-10 mA); while the driving capability of the operational amplifier is relatively small, it can only make the Zener diode work at a smaller operating current. In addition, the dynamic resistance of the Zener diode is larger when working at a small current, which is more likely to cause voltage changes.
(2) Errors introduced by Zener diode temperature drift
The internal structure of 2DW7 (2DW230~236) can be considered as two Zener diodes connected in series. The forward voltage drop of one diode (with a negative temperature coefficient) compensates the temperature drift of the other Zener diode (with a positive temperature coefficient), resulting in a very low temperature coefficient. However, when 2DW7 is applied in reverse, its temperature drift cannot be properly compensated, resulting in a higher temperature coefficient of the negative pulse part.
(3) Error introduced by the operational amplifier
The drift of the input offset voltage of the operational amplifier can directly lead to the error of the pulse amplitude; and if the conversion rate (SR) is too low, it will cause the distortion of the pulse square wave waveform, which will cause voltage error. The price of an operational amplifier with low offset voltage temperature drift and high conversion rate will be very high.
However, for 8~10 bit PWM, this circuit can meet the requirements. For lower requirements, 2 cheap Zener diodes can be connected to replace 2DW7.
3 Application Examples
By using the PWM output of the microcontroller, an RC filter circuit and a first-stage operational amplifier are added to Figure 1 to obtain a 0-10 V DC output voltage as the control signal of the inverter, which has achieved good results. Figure 3 shows an example of using PWM output to control the inverter.
- Benefits of Wide Bandgap Technology for Power Converters
- MT3608 constitutes a 3.7V to 12V boost circuit diagram
- Visible laser digital control modulation driver
- LNK304 non-isolated LED light string driver circuit
- Boost drive circuit composed of RT8450
- Power supply circuit that can reduce LM317 ripple
- Constant current charger using light emitting diode as voltage regulator
- The production of adjustable voltage-stabilized power supply composed of LM317
- Detailed explanation of mobile phone charger circuit diagram
- How to Select Modular EMI Filters for AC-DC Converters
- Block diagram of multi-channel surround stereo power amplifier circuit using digital power amplification
- DC 12V to AC 100V inverter power supply circuit design
- Industrial product promotion and shopping guide DC power supply circuit
- Network card power supply circuit using SC1470 control chip
- MC33060 buck chopper circuit
- 25W single chip audio class D power amplifier IC-TDA7482
- LM629PWM driver
- PWM signal generation circuit diagram
- Push-pull PWM regulated power supply circuit diagram composed of SG3524
- PWM method chopper control circuit for DC motor