BOOSTXL-BUCKCONV: Unexpected results from SFRA on tidu986a

Hi Munadir,

I am looking into this question at the moment. In the meantime, can you refer to our C2000 Software Frequency Response Analyzer (SFRA) Library and Compensation Designer User's Guide? Also, we have a few videos documenting this tool, hope that can clarify some questions you may have

Regards,

Peter

Hi Peter,

Thanks for the suggestion. I had skimmed through the documentation for the SFRA library to try to understand what the Plant and Open-Loop models refer to but the strange results for the open loop buck kind of confused me. Specially I looked at this section: 

My understanding for this would be that for an open loop controller buck converter the frequency analysis in the "Plant-model" and the "Open-Loop" model should be the same or the "Open-Loop" frequency response doesn't really make sense to look at since there is no compensator. 

Anyway, I'll have a more thorough look. 

Hi Munadir,

Did the documentation clear up the questions you had regarding this?

Regards,

Peter

Hi Peter, not quite. The scaling differences were still not evident as to why the DC gain would be lower than expected. However, there is probably a very simple explanation for this that one can probably figure out by digging through the code.

What is more interesting to me and less obvious is the impact of execution frequency and phase rolloff makes no sense. Was there any insight on that?

Hello, 

The expert will be out of the office until 9/2/25. 

Regards,

Carlos

Hi Munadir,

There were slight hardware modifications on the board to facilitate better control loop performance, so the actual SFRA plots will differ slightly from the ones shown in the manual.

Note that roll off is a characteristic that can occur in two directions, depending on the type of filter. For a low-pass filter, the roll-off happens as frequency increases. In a high-pass filter or the lower edge of a band-pass filter, the roll-off occurs as frequency decreases.

The code features a #define for defining the ISR generation divider. Is that what is being used to generate the lower execution frequency?

Regards,

Peter

Hi Peter, I essentially modified the syscfg js file (see attached)

                name: "IsrRatio",
                displayName: "Control Loop ISR Ratio",
                default: 1,
                onChange: recalc_IsrFreq,
                options: [
                    {displayName: "Control ISR runs at Fsw",     name: 1},
                    {displayName: "Control ISR runs at Fsw / 2", name: 2},
                    {displayName: "Control ISR runs at Fsw / 3", name: 3},
		            {displayName: "Control ISR runs at Fsw / 6", name: 6}

Basically the last option "{displayName: "Control ISR runs at Fsw / 6", name: 6}" is what I added. I also wanted to make sure that my changes are working as expected and so I configured a GPIO to toggle everytime the ISR is called.

In "buck_user_setting.h" I added:

#define BUCK_ISR_PIN_TOGGLE_GPIO        4
#define BUCK_ISR_PIN_TOGGLE_CONFIG      GPIO_4_GPIO4

In "buck.h" I added the toggle:

#pragma FUNC_ALWAYS_INLINE(BUCK_runIsr)
static inline void BUCK_runIsr(void)
{
    //
    // Record the ePWM TBCTR value for informational purposes.
    // The value is helpful for measuring the sum of ADC ACQPS + ISR latency.
    // The value can also help to confirm that enough time has elapsed for
    //   the Vout ADC conversion to complete and arrive in ADCxRESULT
    //
    BUCK_isrEnter_ticks = EPWM_getTimeBaseCounterValue(BUCK_DRV_EPWM_BASE);

    GPIO_togglePin(BUCK_ISR_PIN_TOGGLE_GPIO); // Toggle GPIO to indicate ISR rate

In "buck_hal.c" I set the direction and pin configuration:

void BUCK_HAL_setupInterrupt(void)
{
    // Adding handling for ISR pin toggle
    GPIO_setDirectionMode(BUCK_ISR_PIN_TOGGLE_GPIO, GPIO_DIR_MODE_OUT);
    GPIO_setPinConfig(BUCK_ISR_PIN_TOGGLE_CONFIG);

apart from that I didn't make any changes to the project. In my original post I'm measuring GPIO4 to make sure that my ISR is executing as expected and confirmed that my modifications were working like I expected them to both when sysconfig is set to execute ISR at 200kHz and at 200kHz/6 (the setting I added).

Hi Munadir,

That makes sense and it looks like your alterations work. Have you tried any other variations in the control frequency apart from 200kHz and 33kHz? Are the changes in the roll-off gradually transitioning between the 2 frequencies just in the opposite direction you are expecting?

Regards,

Peter

Hi Peter, I have not specifically tried that. Let me give it a try and will keep you posted.