UCD3138: Event setting value for phase shift control

It is counterintuitive, but it is possible to put event 4 at the beginning of the period.  It has some benefits especially if you are using internal or external sync.  It guarantees that if the sync comes before the end of the period, the DPWMB output will still be turned off.  It is also helpful in moving the DPWM2 event out of the DPWM initialization time at the beginning of the period.  It's not good to have an edge moving in and out of this region.  Event 4 is a fixed edge, so it is OK in this area.  

I'm not sure the exact reason we are doing this here, but let me assure that it is legitimate, and often a good idea.  You can still see the dead time by the difference between event 1 and event 4.  

Thank you for your reply.

The figure below is the figure described in the UCD3138 Technical Reference Manual (Literature Number: SNIU028C).
In this figure, the event relationship is "pwmX_ev1 < pwmX_ev2 < pwmX_ev3 < pwmX_ev4".

If you set event 4 at the beginning of the period as in your answer, I think the relationship between the edges of PWMA and PWMB is as shown below. Is this perception correct?

Also, in the sample program described in the first question, pwm0_ev1 and pwm0_ev4 are It will be the same value.
Doesn't this mean there is no dead time?

Best regards,

The first example is also a valid way to use the DPWM and event 4.  It's just not the only one.

In the second case, I don't know what sample software you are referring to, but if you put the same value in Event1 and Event 4, you will get no dead time.

If you put a smaller value in Event4, however, you will get dead time.  

Thank you for your reply.

The second case uses the software (UCD3138FW-PSFB) from TI's UCD3138 website (https://www.ti.com/product/UCD3138#design-development##software-development).

Refer to the derivation formula of the setting value "PWM0_ev4" of event 4 in "deadtime_adjust.c" of the sample program (UCD3138FW-PSFB).

"PWM0_ev4" is calculated using "io_sense" obtained from the ADC.However, the value "sr_falling_dridge_delay" during the calculation is clamped to 0 by the limiter regardless of the value of "io_sense".Therefore, event 4 is actually a fixed value and its value is the same as pwm0_ev1.

A configuration that normally does not cause dead time is not acceptable.

In order to solve this problem, I think that it's necessary to understand the theory of the derivation algorithm of PWM_ev4 in this sample program.

Best regards,

Thank you for your reply.

I didn't have any recognition that I was using DPWM1C and 2C.

I will also investigate how to use DPWM1C and 2C.

Does sending a personal message mean filling out my contact information here? For security reasons(For example, spam mail prevention), I can't publish my contacts in an environment that can be viewed by an unspecified number of people. Also, I'm not allowed to use email addresses other than those specified by my company for business purposes. So I can't even get a temporary email address.

Is there any way I can receive a guide from you other than to filling out my contact information here?

Best regards,

I have sent you a friendship request.

Many Thanks.

Regards,