TMS320F28069M: one Phase sPWM in H-Bridge using F28069M

Yansong,
I may be able to help if you provide some more details on what is happening. What do you mean when you say there is noise when the model is downloaded?

I would suggest that you do not connect this to any high power circuitry until you have verified the PWM output.

You will need some sort of dead band whether you use the Dead-Band sub module or if you adjust your CMPx values to account for the dead band.

Regards,
Cody

Hi Cody,

Thanks for your Reply. The noise after download the model(code) i thought from MOSFETs, the circuit was short circut as the below picture shows. if i didn't push the "Reset" buttom on the board, the board will be Burnout, and must new board to use.

And now i used just 20V as power circuitry. As you said, i have already enabled the Deadband, but it can't realise the function of sPWM in H-Bridge.

So i test the 2 Half-Bridge ePWM4A/B and ePWM5A/B separately with Deadband, and 2 MOSFETS by the Control from ePWM Block operate normally.

After i knew how to use the deadband in order to make the alternating complementary control Signals to ePWM4A and to ePWM4B. For the ePWM5A/B the Control Signal is inverted to ePWM4A/B after synchronization between 2 ePWM blocks. So the source of "RED" and "FED" for ePWM4 are both from ePWM4A, but the source of "RED" and "FED" for ePWM5 are both from ePWM5B. the reason for the choice in order to switch on and off the 4A/5B and 4B/5A in H-Bridge.

 

The Parameter Adjustment of ePWM4 and ePWM5 as below Pic.

But when i connect the OUTPUT1 and OUTPUT2 with resistor as Picture1. there is no Signal between two sides of resistor in oscilloscope. the Signal in Pic.2 cant realise, although theoretically right. so i thought Maybe the  synchronization Problem. so i connect 2 resistors seperatly between OUTPUT1 to GND and OUTPUT2 to GND in oder to see, whether the Signal PWM Output are complementary. but after test, the two Signals from OUTPUT1 and OUTPUT2 are exactly same, so it means, 4A/5A switch on and off at the same time and 4B/5B switch on and off at the same time. but they should not.

i have already tried adjust the Control Signal to gate or a "Phase value" to delay the counter of ePWM5. but i can't figure it out, how can i realise the foundamental function of Pic 2. Is it the Problem of Synchronization? or anything else.

Regards,

Yansong

Yansong,

Forgive me asking a basic question, but why don't you just use one PWM pair for this? Since 4A & 5B are always switched together, and likewise 4B & 5A, just remove PWM5 completely and connect PWM4 in the appropriate places.  Then your dead-band problems would go away.  

Regards,

Richard

Hi Richard,
It's okay, i had also this thought, but it's the work i must to figure it out, and there is no negative Signal at power site. it means on the board, there is no Neutral Point to GND connected. so if i use just one Half-Bridge as Output, so i maybe can't achieve the negative part of sPWM and also i can't get the Sine Wave as output later.
I thought it's maybe not the Dead-band Problem but the synchronization. i'm not sure.
Thanks for your advice.
Regards,
Yansong

Hi venkatesh,
Thanks for your Reply, i'm sure the ePWMs work Right and deadband active.
But sine type as sample based over time based makes Nothing different. There is Always noise after download the Model.
Regards,
Yansong

Yansong,

Will you make the change that Richard suggested and remove the unneeded PWMs?

Can you provide a oscilloscope capture of the 4 PWM signals when there is no load connected? I would like to see the dead-band between these signals.

Regards,
Cody