• State
  • Locked Locked
  • Replies 7 replies
  • Subscribers 82 subscribers
  • Views 146 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TMS320F28386D: Using CLB to XOR PWM signals: Dead-band issue

Hoang Nguyen
Prodigy 70 points

Part Number: TMS320F28386D

Hello everyone,

I am currently using the CLB to XOR three PWM signals. I was able to successfully generate the resulting EPWMxA signal by following Example 12.

However, this approach only works correctly for EPWMxA. When I apply the same XOR operation to the EPWMxB signals, the resulting EPWMxB output does not satisfy the required deadband timing.

Could anyone suggest a proper way to handle this issue or recommend an alternative approach to maintain correct deadband behavior?

Thank you in advance.

  • 0
    TI__Expert 3770 points

    Hi Hoang, 

    Just for clarification, is the issue you are facing that the EPWMxB signals are not properly XORing together within the CLB LUT or is the issue that the outputted correctly XOR'd EPWMxB signal does not have the correct amount of deadband time? 

    Kind regards,
    AJ Favela 

  • 0 in reply to AJ Favela
    Prodigy 70 points

    I mean the outputted correctly XOR'd EPWMxB signal does not have the correct amount of deadband time as I expected. 
    I’m wondering whether there is another way to generate ePWM B while still ensuring proper deadband.

  • 0 in reply to Hoang Nguyen
    TI__Expert 3770 points

    Hi Hoang,

    Can you show me the waveforms of your EPWMxA and EPWMxB for comparison and what the difference in deadband timing is?

    Can you also confirm the configuration within the CLB for both EPWMxB and EPWMxA is the same and if you are feeding these signals back into the EPWM deadband submodule after XORing them?

    Kind regards,
    AJ Favela 

  • 0 in reply to AJ Favela
    Prodigy 70 points

    Hello AJ Favela,

    How can I feed epwmxA signal back into the epwm deadband submodule?

    As I see, the output of CLB is exported to the GPIO by replacing the output of epwmxA as in Example 12. Is there something I misunderstand about output of CLb?

    Thanks.

  • 0 in reply to Hoang Nguyen
    TI__Expert 3770 points

    Hi Hoang,

    I apologize, I was asking if you were attempting to feed the signals back into the dead band submodule after XORing them using the CLB. I see that is not the case. 

    You are correct in your understanding of example 12, the resulted AND'd/XOR'd EPWM signal replaces the output of EPWMxA which is then exported to its GPIO.  

    This is done by selecting the the correct output in the Overriding Outputs selection. To do this for EPWMxA the example uses Output0 which is OUTLUT0. But for EPWMxB you need to use Output 2 which is OUTLUT2, as seen in the table below. Can you confirm that you are using the correct Output to be overwritten? 

                                           

  • 0 in reply to AJ Favela
    Prodigy 70 points

    Hi AJ Favela,

    I know how to use OUTLUT2 to export epwmB. However, cause I xor 3 epwms, the outputs of epwmB and A do not meet deadband requirements.

    So, I just ask for a way to handle this issue or recommend an alternative approach to maintain correct deadband behavior?

  • 0 in reply to Hoang Nguyen
    TI__Expert 3770 points

    Hoang,

    Can you show me a waveform of what your current XOR's EPWM signal looks like and explain what the expected deadband timing behavior should be? 

    You mention that this worked for EPWMxA but not for EPWMxB, what is the difference between these two signals?

    Kind regards,

    AJ Favela