PWM signal in f28049c card using C2000 code generator

Question

Hello,

I am trying to implement a code in my f28049c to control the TI boostxl-buckconv card (a synchronous buck converter). I wanted to do a test asking for two GPIO to input a square wave (pulse generation block) to the mosfets at a frequency of 20 kHz with a duty cycle of 0.5 and discretization period of 5e-6 s. I am measuring the gate signal with an oscilloscope and could not see any signal. If I implement a signal of 10 kHz with a discretization of 1e-5 it is working but every time the discretization is less than 1e-5s I could not get any gating signal. Does anybody have a solution to that or ran through this issue? I need to implement an MPC control algorithm with at least a sampling of 25e-6 s(40 kHz).

 

Best,

T. Harzig

Answer 1

Hi Harzig,

I was not able to implement a step size of 5e-6 seconds (200 kHz) either, I think it is too fast for this controller. However a sample step size of 12.5e-6 seconds (80 kHz) seems to work.

I implemented the attached model on a 280049C launchpad. There are two subsystems. The subsystem labeled "280049_pwm" is using a PWM block instead of the digital outs. The advantage of this implementation is that irrespective of the duty cycle value, you can have a sample step-size same as the switching period.

The subsystem labeled "280049_digital" is using a pulse generator block. The disadvantage here is that you need to consider the duty cycle value in the sample step-size. For example, for a switching frequency of 40e3 and duty cycle of 0.5, the sample step size must be an integer multiple of 0.5/40e3.

If you want to further discuss your controller implementation, email us at support@plexim.com.

Regards,
Manu

Comments

Hello Manu,

Thank you for your answer and for the simulation you just sent me. I implemented my control scheme on the card and I noticed couple issues (with a sampling period of 25 us). When I run the external mode on my laptop I notice that the CPU load block display a value of 80 and then not long after, my laptop is not capable of detecting the card anymore. Is 80 for CPU load too much ?

Also, using a simulation of the synchronous buck I managed to draw what I am supposed to have for the gate signals. When comparing the simulation results to the actual the gate signals I have something way different and it seems that the controller stops working.

You will see in the attached file that i kept the LED blinking in a task and the control scheme in another. I set up the multi tasking mode (to separate the blinking from the actual controller scheme). Can this have an impact on how the gate signals are sent to the MOSFET ?

Sorry for the late answer

Best,

Thibaut Harzig