TMS320F28379D: 3 Phase SPWM

Update:

Using the Trip Zone submodule was able to get the following screenshot.

Using a LAUNCHXL-F38279D. Set up looks like rats nest.  Not sure why so much noise on the positive half cycle. (looking into this)

Also using 2 ePWM blocks.. one for SW1 and a second for SW2.. done because only one TZSEL per ePWM block.

Currently looking for  alternative method:

• Looking for cleaner result
• Looking for way to do 2 spwm's from one ePWM block.

Suggestions welcome.

Best,

Colin

Hi Colin,

Colin Bryant said:

Looking for way to do 2 spwm's from one ePWM block.

You can achieve two sPWMs from one ePWM block by using ePWM1A and ePWM1B. You can look into the dead band submodule to generate active high, low, high complementary, and low complementary signals. 

For the noise, it seems to be coming from the setup, could you provide a picture of your setup? How are you scoping the signals?

Best,

Ryan Ma

Hi Ryan,

Thanks for the input and suggestions.

Will work on sending picture tomorrow.

Using a 3 output sig gen for 3 phase input signals; so these are coaxial connections to a breadboard..

Then jumper'ing breadboard wires to LAUNCHXL board header pins.  (We could get rid of the breadboard)  Am going to try and clean up and minimize wire lengths and connections soon.

Scoping... again using the breadboard wires connected to the header pins of interest.  At one point was probing Launchxl pins directly but was worried about shorting header pins.  Picking up gnd from the breadboard, probably not the best.. prone to gnd loop issues.

Breadboard wires image: 

Will explore the Deadband submodule.

Once we clean up the connections and get a tighter gnd will remeasure stuff.

Best,

Colin

Hi Colin,

I will leave this thread open then and await your updates.

Best,

Ryan Ma

Hello Ryan,

Below is image of the rats nest, got some more appropriate and shorter wires recently.. Plan on minimizing and re-wiring soon.

Colin

Hi Colin,

Thank you for the update, waiting on the minimization of the wires to reduce noise.

Best,

Ryan Ma

Have another related question.

Am now trying to apply some logic gates to the output of the ePWM modules within the MCU and using a digital out GPIO to view the result.

Referencing the images below..

The yellow signal is using the MCU ePWM module driving actual IC logic gates.

The green signal is using the MCU ePWM output an applying logic within the MCU then using dig out GPIO to view result.

Yellow is the desired signal and green is a non-desired result.   

If a GPIO can output a nice looking epwm from epwm module.. what would we have to do to get a nice signal post some logic operations on ePWM output to GPIO port. 

right hand side of image above.

Best,

Colin

Hi Colin,

Colin Bryant said:

The green signal is using the MCU ePWM output an applying logic within the MCU then using dig out GPIO to view result.

How are you applying the "logic" within the MCU? Are you using ePWM module for this or another peripheral? What logic are you trying to apply?

Is the yellow signal without the internal logic being applied? If so, is the ePWM module behaving as expected? 

Best,

Ryan Ma

Hi Ryan,

Using Matlab/Simulink to generate the code for LAUNCHXL-F28379.  So we are building a model in Simulink then using embedded coder to program MCU.

For our application we need to apply some logic to the ePWM outputs to get the pwm's needed.  Not sure we have a good grip on how to use the total functionality of the ePWM blocks but it would be really cool if we could apply logic between the ePWM outputs within the ePWM module.. example.. "ePWM1A && ePWM1B" as final output from ePWM1.  Our application... we are "and'ing" with some "or'ing" operations.

For the green signal ("not so good signal") we are taking the MCU ePWM GPIO outputs and connecting them back into MCU via GPIO digital inputs then applying "logical operator" blocks from Simulink.  The output of logical operation is then sent to a digital output GPIO to see result. I am not sure if issue is with digital GPIO output or logic operators, leaning towards the GPIO output or the way it is configured.

The yellow signal ("the good signal") was result of taking the MCU ePWM GPIO outputs and connecting them directly to external logic IC's on a breadboard, same logic operation as green signal, then observing it's output. This produced good and expected output of applying logical operations to ePWM outputs, although externally done.

Best,

Colin

Hi Ryan,

Sure that will work.  Thanks for your help!

Best,

Colin