Dead Zone Problem in PWM Modulation

　　The concept of "dead zone" must be recorded, collected from the Internet, and used when it is useful.

　　PWM pulse width modulation is most commonly used in power electronics for rectification and inversion. This requires the use of rectifier bridges and inverter bridges. Taking two-level as an example, there are two power electronic devices on each bridge arm, such as IGBTs. These two IGBTs cannot be turned on at the same time, otherwise a short circuit will occur. Therefore, designing a PWM wave with a dead zone can prevent the upper and lower devices from being turned on at the same time. In other words, when one device is turned on and then turned off, it will pass through a dead zone before the other device can be turned on.

　　Dead zone, usually, high-power motors, inverters, etc., are all H-bridges or 3-phase bridges composed of high-power tubes, IGBTs and other components at the end. The upper and lower half bridges of each bridge must not be turned on at the same time, but when the high-speed PWM drive signal reaches the control pole of the power component, it often produces a delay effect due to various reasons, causing a half-bridge component to not be turned off when it should be turned off, causing the power component to burn out. The dead zone is to delay for a period of time before turning on the lower half bridge after the upper half bridge is turned off, or to delay for a period of time before turning on the upper half bridge after the lower half bridge is turned off, so as to avoid burning out of the power component. This delay time is the dead zone. (That is, the components of the upper and lower half bridges are both turned off)

　　Dead time control is not available in the PWM equipped by common low-end microcontrollers. Dead time is a protection period set during PWM output to prevent the upper and lower tubes of the H-bridge or half-H-bridge from being turned on at the same time due to switching speed issues. Therefore, during this time, neither the upper nor lower tubes will have output, which will of course interrupt the waveform output. The dead time generally only accounts for a few percent of the cycle. However, when the duty cycle of the PWM wave itself is small, the empty part is larger than the dead time, so the dead time will affect the output ripple, but it should not play a decisive role.

　　The duty cycle is the ratio of the time the high level is maintained in the output PWM to the time of the PWM clock cycle; for example, if the frequency of a PWM is 1000Hz, then its clock cycle is 1ms, which is 1000us. If the high level appears for 200us, then the low level must be 800us, so the duty cycle is 200:1000, which means that the PWM duty cycle is 1:5.

　　Resolution is the minimum duty cycle that can be achieved. For example, the theoretical resolution of 8-bit PWM is 1:255 (single slope), and the theoretical resolution of 16-bit PWM is 1:65535 (single slope). The frequency is like this. For example, the resolution of 16-bit PWM reaches 1:65535. To achieve this resolution, T/C must count from 0 to 65535. If the count is from 0 to 80 and then from 0 to 80, then its minimum resolution is 1:80, but it is also faster, which means that the output frequency of PWM is high.

　　Suppose a PWM counts from 0 to 80, and then counts from 0 to 80 again. This is a single slope. Dual slope / Single slope Suppose a PWM counts from 0 to 80, and then counts from 80 to 0. This is a dual slope. It can be seen that the counting time of the dual slope is doubled, so the output PWM frequency is half as slow, but the resolution is 1: (80 + 80) = 1: 160, which is doubled. Assuming that PWM is a single slope, the maximum count is set to 80, and we set a comparison value of 10, then when T/C counts from 0 to 10 (at this time, the counter still counts up until the count reaches the set value 80), the microcontroller will control a certain IO port to output 1 or output 0 or port inversion at this time according to your settings. This is the most basic principle of PWM.