LAUNCHXL-F280049C: CMPSS trip to cause three different configurations of the EPMW AQ

Hi Anurag,

Just want to clarify: are you trying to have the EPWM react in 3 different ways but sourced from a single CMPSS signal? Or are you using three separate CMPSS instances?

If you are using CMPSS, the signal path would be to use CMPSS TRIPH or TRIPL routed through EPWM XBAR and then use the Digital Compare and/or Trip Zone submodules to change EPWM outputs - a good place to look at this would be the TRM Digital Compare Section (specifically, Figure 18-49. Digital-Compare Submodule High-Level Block Diagram). You could also use T1 and T2 of the Action Qualifier submodule and source these events from the comparator, but I believe you are limited to 2 configuration options with that and not 3. Was there a specific submodule you were targeting to use for this? 

Best Regards,

Allison

I'm using three different CMPSS instances. I was thinking of following -

1. The EPWM1A/B (complementary pairs) are triggered in following way -

CMPSS1H --> TRIP4 --> DCAEVT1 --> T1 --> AQ goes HIGH

CMPSS1L --> TRIP5 --> DCAEVT2 --> T2 --> AQ goes LOW

2. The EPWM2A/B (complementary pairs) are triggered in following way -

CMPSS1H --> TRIP4 --> DCAEVT1 --> T1 --> AQ goes HIGH

CMPSS7H --> TRIP7 --> DCAEVT2 --> T2 --> AQ goes LOW

Is this possible to do?

Also I would like to change the configuration of which CMPSS trip signal triggers which AQ in real time based on some mode switching. Is that also possible?

On a second note I think that the alternate implementation could be using a CLB because CMPSS triggers can be taken to the CLB and the I have logic expression which gives the output gating signals. Any advice on which approach would be cleaner and would have more execution efficiency?

Hi Anurag,

Items 1 and 2 look like fine configurations to use T1 and T2 - and, yes, you could change which CMPSS trip is used in software. I don't think there would be timing issues if you are updating the action qualifiers - when would you be updating this? I also was able to discuss with a few others and it seems like not using CLB would be cleaner. Let me know if you have further questions/issues with this!

Best Regards,

Allison

Hey Allison,

I would be updating them if the mode of my controller changes based on some user input. That's very slow change compared to switching frequency. Can I know why CLB is not the way to go?

Hi Anurag,

Allison is currently out of office, but you can expect a response from her on Monday 1/29.

Best Regards,

Delaney

Hi Anurag, 

Here is input from a CLB expert:

This should also be possible with the CLB; CLB examples 3 and 4 describe some use cases for the ePWM with CLB (example 3 creates an auxiliary PWM, which may not be what you want). If the action qualifier is just controlling what the ePWM does for its output, it should be possible since the CLB can override certain peripheral outputs like the ePWM. Or if you are just looking to have a signal go low or high depending on certain triggers, then CLB is probably more useful then the ePWM if you don’t need a square wave. But we don't know enough about your particular application to say for sure. Are you trying or prefer to implement CLB for this?

Best Regards,

Allison