TMS320F28375S: TZCTL - Are Trip Zone Configurations Redundant?

Jeff,

The TZx signals have priority over the DCxEVTy signals, and their relationship can be somewhat confusing.

1. If a DC event is setup as a TZ event, and the DC event forces the PWM low and the TZ forces the PWM high, then the TZ event will take priority.
2. If a DC event is setup as a TZ event, and the DC event forces the PWM low and the TZ is set to “Do Nothing”, then the DC even should take effect.

Cody,

That all makes sense.  I'm pretty sure I have it configured for scenario 2 that you outlined; here is the code used to do that.

{
    EALLOW;

    /* Clock config */
    ClkCfgRegs.PERCLKDIVSEL.bit.EPWMCLKDIV = 0u; /* EPWMCLK = PLLSYSCLK / 1 = 200MHz */
    EPwm1Regs.TBCTL.bit.CLKDIV = 7u;    /* Set CLKDIV = 8 */ /* PWM1 CLK = 111607Hz */
    EPwm1Regs.TBCTL.bit.HSPCLKDIV = 7u; /* Set HSPCLKDIV = 1 */

    /** - Clear counter */
    EPwm1Regs.TBCTR = 0u;

    /** - Set control actions */
    EPwm1Regs.AQCTLA.bit.CAU = PWM_ACTION_CLEAR;
    EPwm1Regs.AQCTLA.bit.ZRO = PWM_ACTION_SET;

    /** - Set period for 1ms */
    EPwm1Regs.TBPRD = 700u;
    EPwm1Regs.CMPA.bit.CMPA = 350u;     /* Set duty cycle to 100% */

    /** - EPWM XBar config for TRIP4 */
    EPwmXbarRegs.TRIP4MUX0TO15CFG.bit.MUX7 = 1u;    /* Select INPUTXBAR4 (MUX7.1) */
    EPwmXbarRegs.TRIP4MUXENABLE.bit.MUX2 = 0u;      /* Disable MUX2 */
    EPwmXbarRegs.TRIP4MUXENABLE.bit.MUX7 = 1u;      /* Enable MUX7 */

    /** - Digital Compare A config */
    EPwm1Regs.DCTRIPSEL.bit.DCAHCOMPSEL = 3u;       /* Select TRIP4 as DCAH source */
    EPwm1Regs.DCAHTRIPSEL.bit.TRIPINPUT4 = 1u;      /* Enable TRIP4 as trigger */
    EPwm1Regs.TZDCSEL.bit.DCAEVT1 = 2u;             /* Trigger on DCAH high */

    /** - Select DCAEVT1 as oneshot source */
    EPwm1Regs.TZSEL.bit.DCAEVT1 = 1u;

    /** - Configure input Xbar to read from OV pin */
    InputXbarRegs.INPUT4SELECT = PWM_DEMO_PIN_OV;

    /** - Configure DCFWINDOW for 100us blanking */
    EPwm1Regs.DCFWINDOW = 11u;                      /* ~98.56us but whatevs */
    EPwm1Regs.DCFOFFSET = 0u;
    EPwm1Regs.DCACTL.bit.EVT1SRCSEL = 1u;           /* Select DCEVTFILT as source for DCAEVT1 */
    EPwm1Regs.DCFCTL.bit.PULSESEL = 1u;             /* Align window to TBCTR = 0 */
    EPwm1Regs.DCFCTL.bit.SRCSEL = 0u;               /* Select DCAEVT1 to filter */
    EPwm1Regs.DCFCTL.bit.BLANKINV = 0u;
    EPwm1Regs.DCFCTL.bit.BLANKE = 1u;               /* Enable blanking */

    /** - Enable ePWM1A */
    EPwm1Regs.TBCTR = 0u;
    EPwm1Regs.TBCTL.bit.CTRMODE = 0u;

    /** - Configure trip control actions */
    EPwm1Regs.TZCTL.bit.TZA = 3u;
    EPwm1Regs.TZCTL.bit.DCAEVT1 = 2u;
    EDIS;
}

The only additional detail is that in this configuration I'm trying to create a blanking window, and so really the source of the trip is DCAEVTFLT, not just DCAEVT1.  With this code, no trip action is taken, but in TZFLG the bits DCAEVT1 and OST both indicate a trip occurred.

Is there a flaw in the code I used above?

Jeff,

can you highlight the path that you are trying to setup on the trip-zone submodule diagram below?

Also I would recommend that you disable the DC event filter until you verify that you have everything setup correctly. You could be checking for trigger events that have been filtered, even if the window is only:

Jeff Campbell91 said:

/* ~98.56us but whatevs */

Regards,
Cody

Cody,

The path I'm trying to activate would be as follows:

To that end, this is how I've configured the code (removing filtering for now as per your suggestion):

        /* Section: GPIO Config */

        /** - Configure OV switch pin */
        GPIO_SetupPinMux(PWM_DEMO_PIN_OV, GPIO_MUX_CPU1, 0u);
        GPIO_SetupPinOptions(PWM_DEMO_PIN_OV, GPIO_INPUT, GPIO_SYNC);

        /** PWM Setup: */

        /** - Initialize GPIO for ePWM1A */
        InitEPwm1Gpio();

        /* Section: PWM config */

        EALLOW;

        /** - Set ePWM clock divider */
        ClkCfgRegs.PERCLKDIVSEL.bit.EPWMCLKDIV = 0u; /* EPWMCLK = PLLSYSCLK / 1 = 200MHz */

        /** - Set clock dividers and count mode */
        EPwm1Regs.TBCTL.bit.CLKDIV = 7u;    /* Set CLKDIV = 8 */ /* PWM1 CLK = 111607Hz */
        EPwm1Regs.TBCTL.bit.HSPCLKDIV = 7u; /* Set HSPCLKDIV = 1 */

        /** - Clear counter */
        EPwm1Regs.TBCTR = 0u;

        /** - Set control actions (full on) */
        EPwm1Regs.AQCTLA.bit.CAU = PWM_ACTION_CLEAR;
        EPwm1Regs.AQCTLA.bit.ZRO = PWM_ACTION_SET;

        /** - Set period for 1ms */
        EPwm1Regs.TBPRD = 700u;
        EPwm1Regs.CMPA.bit.CMPA = 350u;     /* Set duty cycle to 50% */

        /* Section: Trip config */

        /** - Configure input Xbar to read from OV pin */
        InputXbarRegs.INPUT4SELECT = PWM_DEMO_PIN_OV;

        /** - EPWM XBar config for TRIP4 */
        EPwmXbarRegs.TRIP4MUX0TO15CFG.bit.MUX7 = 1u;    /* Select INPUTXBAR4 (MUX7.1) */
        EPwmXbarRegs.TRIP4MUXENABLE.bit.MUX7 = 1u;      /* Enable MUX7 */

        /** - Digital Compare A config */
        EPwm1Regs.DCTRIPSEL.bit.DCAHCOMPSEL = 3u;       /* Select TRIP4 as DCAH source */
        EPwm1Regs.DCAHTRIPSEL.bit.TRIPINPUT4 = 1u;      /* Enable TRIP4 as trigger */
        EPwm1Regs.TZDCSEL.bit.DCAEVT1 = 2u;             /* Trigger on DCAH high */

        /** - Select DCAEVT1 as oneshot source */
        EPwm1Regs.TZSEL.bit.DCAEVT1 = 1u;


        /** - Configure trip control actions */
        EPwm1Regs.TZCTL.bit.TZA = 3u;                   /* Ignore TZA */
        EPwm1Regs.TZCTL.bit.DCAEVT1 = 2u;               /* Force low on DCAEVT1 */

        /** - Enable ePWM1A */
        EPwm1Regs.TBCTR = 0u;
        EPwm1Regs.TBCTL.bit.CTRMODE = 0u;

        EDIS;

With this configuration, I can observe the trip event DCAEVT1 firing, but no effect on the PWM signal.  I can fix this by setting TZCTL[TZA] to 2 (force low) or 0 (HiZ mode), but it seems like this should be able to work.

Cody,

That clears things up significantly. I disabled the OSHT logic entirely and it started to work like I'd expect. You're right, it was that dual-possibility routing that had me confused.

Thanks for the help!
Jeff

Wonderful, glad I could help!

Do you think it would help other customers if I go and add in some extra verbiage around the trip-zone logic?

And of course, if you have any other questions feel free to start another thread.

Regards,
Cody

I think it would help, yes. It was difficult to see on my first read that the DC events could be used directly or as inputs to TZA/B; it looks like that was only hinted at by the block diagram.