TMS470MF03107: PWM signals' period(& duty) generation based on HET displayed on Oscilloscope screen is very different from period(& duty) set in CCS project files generated from HALCogen.

tms470mf03107 HET module pin description:

PCB 4 PWM pin layout:

After ZOOM in：

HET Global Timing Configuration is as follows：

The main function is as follows:

int main(void)

{

/* USER CODE BEGIN (2) */

    hetInit();

            pwmSetSignal(pwm0,set_pwm0);

            pwmSetSignal(pwm1,set_pwm1);

            pwmSetSignal(pwm2,set_pwm2);

            pwmSetSignal(pwm3,set_pwm3);

           // pwmSetSignal(pwm4,set_pwm4);

           // pwmSetSignal(pwm5,set_pwm5);

           // pwmSetSignal(pwm6,set_pwm6);

           // pwmSetSignal(pwm7,set_pwm7);

            pwmStart(pwm0);

            pwmStart(pwm1);

            pwmStart(pwm2);

            pwmStart(pwm3);

           // pwmStart(pwm4);

           // pwmStart(pwm5);

           // pwmStart(pwm6);

          //  pwmStart(pwm7);

            pulse_pwm0 = pwmGetSignal(pwm0);

            pulse_pwm1 = pwmGetSignal(pwm1);

            pulse_pwm2 = pwmGetSignal(pwm2);

            pulse_pwm3 = pwmGetSignal(pwm3);

          //  pulse_pwm4 = pwmGetSignal(pwm4);

          //  pulse_pwm5 = pwmGetSignal(pwm5);

           // pulse_pwm6 = pwmGetSignal(pwm6);

           // pulse_pwm7 = pwmGetSignal(pwm7);

//        }

/* USER CODE END */ 

}

4 PWM signal Period set is the same with the PWM signals generated by program,covering the minimum(482us) and maximum period(1736044us) i want.

BUT the real PWM signal waves are as follows:

S1(PWM0,het[11]):

Constantly HIGH level：

S2(PWM1,het[12]):

displayed PWM wave output period is 290us，which is HALF of the set period 578us：

S3(PWM2,het[13])：

displayed PWM wave output period is 720ms，which is HALF of the set period 1446703us：

S4(PWM3,het[14])：

displayed PWM wave output period is 880ms，which is HALF of the set period 1736044us

When the global timing configuration is as follows,which is default setting(other configuration stays the same):

PWM0 and PWM1 period set is the same with the program-generated value,but program-generated value of PWM2 and PWM3 duty&period are abnormal compared with the duty&period set.You can see as follows:

Oscilloscope Screen displays are as follows:

S1:period=960us,TWICE of the set period 480us

S2:period=1160us,TWICE if the set period 578us

S3:almost constant level 376ms TWICE of 188414us

S4:duty abnormal, constant HIGH level:

Here in order for simplification, some other test and description is as follows：

HET Global Timing Configuration ONE:

4 PWM signal Period set is 10000us(10ms),which is the same with the PWM signals generated：  

But the real PWM pulse waves are as follows:

S1(PWM0,het[11])：NO PWM wave shape displayed on Oscilloscope screen,but pulled HIGH constantly.

S2(PWM1,het[12])，S3(PWM2,het[13])，S4(PWM3,het[14]):PWM waves are displayed as follows on screen,but pulse period(5000us or 5ms) is HALF of the set period(10000us or 10ms)：

WHEN the global timing configuration TWO is as follows:

4 PWM signal Period set is 10000us(10ms),which is the same with the PWM signals generated：

But the real 4 PWM pulse waves are as follows，all having high and low level apparently,but pulse period(20000us or 20ms) is TWICE of the set period(10000us or 10ms).

All in all, I wonder， based on the certain HET timing configuration， why the generated  PWM periods are not the same with the PWM period(& duty) set by me. Please make an answer as in detail as possible.

THX~! 

Cherry Zhou：

Hi,

We're looking this issue and please expect the response by next business day.

Thanks

,

Joseph Chen：

OK then!!THX 

,

Joseph Chen：

Hi Cherry, I hope it's a NICE day for you and your fellows!

For this business day， what do you see about this issue?

Urgently waiting for your response in detail, THX~!

,

Cherry Zhou：

Hi,

Because of the Thanksgiving holiday in the U.S., TI E2ETm design support forum responses may be delayed the week of Nov. 20. Thank you for your patience

Sorry for any inconvenience.

Thanks

,

Joseph Chen：

Any progresses for this issue plz after Thanksgiving holiday???

请问一下，感恩节假期结束了吧？这个问题进展如何了呀？？？

,

Cherry Zhou：

Hi,

Sorry for the late response.

 What is the OSC frequency used by the customer and the rest of clock parameters? Can you show the settings for below?

Thanks

,

Joseph Chen：

Hi Cherry,

       Roger that well! thx.

       As you mentioned,the OSC frequency used by the customer is 8MHz.

       And the rest of the clock parameters is as follows, which is default setting:

    

,

Joseph Chen：

After I set the OSC freq. as 8MHz as follows,the issue still remains:

,

Cherry Zhou：

Hi,

  I don't think the OSC for 8Mhz is really taken by the tool as you can see the OSC still shows as 16Mhz in your screenshots. See below. 

I think after you enter 8Mhz in the Oscillator tab, you need to click the 'GCM' button for the 8Mhz to take effect. 

Thanks and regards,

Cherry

,

Joseph Chen：

Hi,

    My latest  comment to you is already as follows:

    

And HET Global timing is as follows，and  the issue still remains:

,

Cherry Zhou：

Hi,

 Can you try as follows?

Thanks

,

Joseph Chen：

Hi,

     I have tried the possible configuration you offer:

I wonder what is your consideration for this configuration, but the debugging result is also not optimistic~, which you can see as follows with pwm duty and period output not I set:

And the pwm real wave form is as follows, NO PWM wave with dutu and period I want:

Would you please think twice to make a further reply to solve my problem, looking forward urgently, thanks!

,

Cherry Zhou：

Hi,

I'll check with my college and give you the updates ASAP.

Thanks

,

Cherry Zhou：

Hi,

This device is very old and I have not worked on this device before. Reading the TRM, it looks like VCLK2 cannot be slower than VCLK. Therefore, please also change VCLK to 20Mhz the same as VCLK2. 

Thanks

,

Joseph Chen：

Hi 

Thx for your answer,~!

After I tried the way you offer, the result is as follows, which is still not optimistic: 

S1:

S2:

S3:

S4:

,

Cherry Zhou：

Hi,

Please allow some time for us to get back to you.

Thanks

,

Cherry Zhou：

Hi,

Please refer to Time Base section 12.3.3 in the TRM about HR and LR configuration. You can't generate a PWM with a period that is smaller than a LR. 

Thanks

,

Joseph Chen：

Hi,

As you said, the LR we put it before is 3200ns, that is 3.2us, and my minimum PWM period set is 480us, which is much larger than LR ~~~

According to your logic, what HR & LR etc. parameters configuration can satisfied the desired PWM period generation ???  

,

Cherry Zhou：

Hi,

Sorry, I don't know why three other PWM works but only the first one does not (50% duty with 480uS period). Can you tried a few things?

  – At what PWM frequency does pwm0 work? 

  – I want to know if there is something special about pwn0. Can you do an experiment. Configure pwm0 with 50% duty and 578uS just like pwm1? Does it produce the PWM?

  – For pwm1, pwm2 and pwm3, can you try 50% duty and 480uS. I want to know if all pwm will fail at this frequency.

  PLEASE be aware of the long holiday next week in the US. The response will be extremely delayed.  

(Please see the link here for more clarification)

Sorry for any inconvenience.

Thanks