Deadtime about PWM

[Jerry Yung]

i config TIM with PWM , and insert deadtime , i found the first deadtime is longer than later , this why behave so ?

​

[waclawek.jan]

What is exactly the setup (in terms of TIM registers, I don't understand Cube/HAL), and what waveform corresponds to which pin?

JW

[Jerry Yung]

pin: CH1N CH1

CCMR: OC1M = 0x06​

CCER: CC1E = 1 ,CC1NE = 1 ,CC1P =0,CC1NP = 1

BDTR: DTG = 0xC8

that's all .

[waclawek.jan]

If you use PWM mode 1 with CC1P=0 and CC1NP=1, and from reset state set up first CCR1, CCMR1, CCER, ARR, both pins should go high, in timing depending on the order of how these settings are performed, as the counter is not running yet, CNT == 0, and in PWM1mode, while CNT<CCR1, the non-inverted output is active.

In other words, the first two rising edges are most probably artefact from the setup process. You may try to set up everything except CCER, and then enable both channels in a single write of CCER, before starting the counter (CR1.CEN=1). You should then see the first two edges to rise simultaneously, and then the rest of the waveform as expected.

JW

[Jerry Yung]

yes, the first two edges rise ​simultaneously,

but I want a deadtime between the firse two edges

[waclawek.jan]

Try to set CNT to ARR-1 just before enabling the timer.

JW