TMS320F280049: Timing delay until disable PWM

Kengo Furuya

Part Number: TMS320F280049
Other Parts Discussed in Thread: C2000WARE

Hello C2000,

My customer see the timing variation until disable PWM from receive the error signal of F280049. Sometime it take a 9.2usec and some time 3.6msec.

They are using CMPSS to trip the ePWM. The flow is below.

CMPSS receive the analog error signal and output interrupt signal to TZ of ePWM.

After that, through ePIE, software disable the PWM output by AQCSFRC.

Why does such kind of timing variations occur? Please let me know the possible cause.

Regards,

Furuya

over 7 years ago

0 Frank Ankapong over 7 years ago

TI__Mastermind 28275 points

Hi Kengo,

Can you add a little more detail? Are you saying the time from a CMPSS trip to EPWM shutdown is varying from 9.2us to 3.6ms?

0 Kengo Furuya over 7 years ago in reply to Frank Ankapong

TI__Expert 7650 points

Yes, your understanding correct. Customer see the time variation from CMPSS trip to EPWM shutdown.

Regards,

Furuya

0 Frank Ankapong over 7 years ago in reply to Kengo Furuya

TI__Mastermind 28275 points

Ok. That shouldn't be the case. Can you try the C2000Ware example "cmpss_ex1_asynch"?

0 Frank Ankapong over 7 years ago in reply to Frank Ankapong

TI__Mastermind 28275 points

Hi Kengo,

I looked through the additional details you supplied by email and i think the variation you are seeing is coming from the fact that you are shutting down the ePWM using AQCSFRC. Since you are doing this in software and in an interrupt, the variation is not unreasonable. Unless you need to do some post-processing once the cmpss trip occurs, there is no need to approach it this way. The functionality you are looking for "cmpss trip to ePWM shutdown" can be completely done in hardware. You can program the TZ module of the ePWM to force the output of the ePWM to a certain state once a trip condition occurs from the CMPSS. This will happen in hardware and it will actually end up being much faster, in the nanoseconds. Please look at the "cmpss_ex1_asynch" example again because that is what it's doing.

0 Kengo Furuya over 7 years ago in reply to Frank Ankapong

TI__Expert 7650 points

Hi Frank,

Customer understand the cause might be disabling PWM by software based on interrupt.
But, now they need to disable the output of PWM synchronously with PWM cycle. So they use AQCSFRC for disabling the synchronization.
If there are any method to achieve the same function by using CMPSS and ePWM trip, could you let me know how to set it?

Regards,
Furuya

0 Frank Ankapong over 7 years ago in reply to Kengo Furuya

TI__Mastermind 28275 points

Hi Kengo,

Yes, you can still use the TZ module and configure it to do a cbc (cycle-by-cycle trip). For cbc, the trip condition coming into the TZ submodule from the DC submodule is re-evaluated for every cycle. From your description, I'm guessing that's what you are looking for. If that isn't what you are looking for, can you provide a diagram description of the functionality you want?

0 Frank Ankapong over 7 years ago in reply to Frank Ankapong

TI__Mastermind 28275 points

Hi Kengo,

Was the customer able to resolve the issue?

0 Frank Ankapong over 7 years ago in reply to Frank Ankapong

TI__Mastermind 28275 points

Hi Kengo,

It's been a while since we heard from you so I'm going to assume the suggestion offered earlier resolved the issues. I'll be marking this thread as closed. If that is not the case, you can reply to this thread or create a new one if this one locks out due to inactivity.

0 Kengo Furuya over 7 years ago in reply to Frank Ankapong

TI__Expert 7650 points

Frank,

Now customer is confirming if they can use TZ module with CBC. If any feedback, I will re-open this thread.

Regards,
Furuya

0 Frank Ankapong over 7 years ago in reply to Kengo Furuya

TI__Mastermind 28275 points

Sure Kengo, I'll mark this as closed for now.

C2000™︎ microcontrollers

C2000 microcontrollers forum

TMS320F280049: Timing delay until disable PWM