CCS/TMS320F28075: Problem with ePWM X-BAR

Jong Woo Kim

Part Number: TMS320F28075
Other Parts Discussed in Thread: CONTROLSUITE, C2000WARE

Tool/software: Code Composer Studio

Hello, I am Jongwoo Kim, and I am struggling with ePWM X-BAR.

My purpose is to produce DCAEVT1.sync signal from CMPSS.CTRIPH so that an ePWM module can be synchronized with the rising edge of the comparator output.

I tested 3 ways to achieve this function as follows:

1. CMPSS.CTRIPOUTH -> GPIO_OUT ------ wired connection ----> GPIO_in -> EXTSYNCIN -> synchronized.

2. CMPSS.CTRIPOUTH -> GPIO_OUT ------ wired connection ----> GPIO_in -> GPIO X-BAR -> TRIPx -> DCAH -> DCAEVT1.sync -> synchronized.

3. CMPSS.CTRPH -> ePWM X-BAR -> TRIPx -> DCAH -> DCAEVT1 .sync -> synchronized.

I successfully got the synchronized ePWM by using the methods 1 and 2. Therefore, I think there is no problem in CMPSS and DC module.

However, when I tried the method 3 (this is the most attractive way since it does not use the connection outside), ePWM module is not synchronized. just free running.

Therefore, I think the CMPSS.CTRPH -> ePWM X-BAR -> TRIPx steps in the method 3 has a problem.

In this case, could you give me some help? The followings are my settings for CMPSS and ePWM module. (please note that the comparator output and DC submodule are fine as I mentioned before)

(Comparator output is okay. CTRIPOUTH is also okay. But I am not sure CTRIPH is okay.)

Cmpss2Regs.COMPCTL.bit.CTRIPHSEL = 2; // Digital filtered comparator output is uploaded to ePWM X-BAR

EPwmXbarRegs.TRIP10MUX0TO15CFG.bit.MUX2 = 1; // CMPSS2.CTRIPH -> MUX2 output -> TRIP10MUXENABLE switch

EPwmXbarRegs.TRIP10MUXENABLE.bit.MUX2 = 1; // TRIP10MUXENABLE enables MUX2

EPwmXbarRegs.TRIPOUTINV.bit.TRIP10 = 0; // Active high

EPwm7Regs.DCTRIPSEL.bit.DCAHCOMPSEL = 9; // TRIPIN10 is selected for DCAH

(I also tested DC submodule so that DCAH->DCAEVT1.sync is working.)

Thank you for your interest in advance.

Sincerely,

over 8 years ago

0 Vivek Singh over 8 years ago

TI__Guru** 115271 points

Hi Kim,

We'll look into your query on EPWM XBAR.

For method 1 & 2, , you don't have to make any external connection. GPIO_OUT will be looped back to same GPIO_IN and you can configure the INPUT XBAR to connect this GPIO_IN to EXTSYNCINx signal (please refer Figure 8-1. Input X-BAR in TRM). Can you try this?

Regards,

Vivek Singh

0 Manoj Santha Mohan over 8 years ago

TI__Guru 55700 points

Kim,

Did you configure TZDCSEL register to generate DCAEVT1/2 signal?
Did you enable EVTxSYNCE to generate sync signal based of DCAEVTx signals

Regards,
Manoj

0 Jong Woo Kim over 8 years ago in reply to Vivek Singh

Prodigy 120 points

Hi Vivek,

Thank you for your interest and suggestion. I didn't know that an GPIO_OUT can be directly sent to INPUT XBAR. I will try this and share the result.

Yours sincerely,

Jongwoo Kim

0 Jong Woo Kim over 8 years ago in reply to Manoj Santha Mohan

Prodigy 120 points

Hi Manoj,

As I metioned, I configured and enabled TZDCSEL and EVTxSYNCE. I also tested DC submodule so that DCAH->DCAEVT1.sync is working with the exteranl connections by INPUT XBAR with TRIP6. The problem was that I cannot use ePWM XBAR for the trigger of DCAH. Thank you for your interest!

Yours sincerely,

Jongwoo Kim

0 Manoj Santha Mohan over 8 years ago

TI__Guru 55700 points

Did you make sure CTRIPOUTH signal is generated? Please check XBARFLG register to see whether you received the comparator subsystem trip signal.

I would also try forcing the trip signal.

-Manoj

0 Jong Woo Kim over 8 years ago in reply to Vivek Singh

Prodigy 120 points

Hi vivek,

The GPIO assigned COMPOUT signal can go back into the microcontroller by just assigning the pin to the INPUT XBAR.
I got the syncronized signal without connecting the wire outside.
However, I need to assign one GPIO pin in order do obtain this function.
So, could you check whether ePWM XBAR can be used for this function so that I don't need to assign a GPIO pin?

Thank you so much!

Sincerely,

Jongwoo Kim

0 Jong Woo Kim over 8 years ago in reply to Manoj Santha Mohan

Prodigy 120 points

Manoj,

Sure. I checked CTRIPOUH signal from the GPIO pin. Also, synchronizing function is also checked by feeding CTRIPOUTH back to INPUT XBAR and TRIP6. I was not able to use ePWM XBAR for this function, although using ePWM XBAR is the shortest way in this structure.
Thank you so much!

Sincerely,

Jongwoo Kim

0 Manoj Santha Mohan over 8 years ago in reply to Jong Woo Kim

TI__Guru 55700 points

Jong,

You should be able to use ePWM XBAR approach. Little bit of debug should get you there. But, if you prefer to use method 1/2. Thats also fine.

Regards,
Manoj

0 tlee over 8 years ago in reply to Manoj Santha Mohan

TI__Guru 62975 points

Jong,

Were you able to try the "cmpss_asynch" example in controlSUITE or C2000Ware? That should be an example of configuring the CMPSS to EPWM trip path.

Also, be aware that MUX2=1 will pass through the OR'd value of CTRIPH + CTRIPL, whereas MUX2=0 will pass through CTRIPH only: EPwmXbarRegs.TRIP10MUX0TO15CFG.bit.MUX2 = 1; // CMPSS2.CTRIPH -> MUX2 output -> TRIP10MUXENABLE switch

-Tommy

0 Jong Woo Kim over 8 years ago in reply to tlee

Prodigy 120 points

Tommy,

I will try to look into the example. I was doing myself.

Also, I think there could be some typo. What I saw from the table 7-2 was CMPSS1.CTRLH is assigned to MUX0 - 1.

However, in Figure 7-2, each mux input is allocated from zero to 3.

I was wondering this point but forgot to test MUX0-0. I will try this method, and I think this could solve this problem.

I will share the result.

Thank you.

Sincerely,

Jongwoo

0 Jong Woo Kim over 8 years ago in reply to tlee

Prodigy 120 points

Tommy,

I just found that MUX2=0 works..
Please note that table 7-2 contains the wrong mux information. It should be started from zero.
I appreciate your solution.

Sincerely,

Jongwoo