TMS320F28034: Question about eSMO in DMC MATH_eSMOlib_full.pdf

Jack Tan80

Dear C2000 expert,

From DMC MATH_eSMOlib_full.pdf, I have below 2 questions, can you please help on this?

1. I could understand the equation 1 in the below picture, but in equation 2, there is an extra variable z, can you please help me understand how does the z comes from?

2. The equation 4 and 5 are discrete form of equations 2 and 3, the parameter I derived should be F = (1+ATs), G = BTs. Can you please help me know why?

My derivation is like this, for equation 2, the discrete form should be, i(n+1) - i(n) = ATs*i(n) + BTs(vs(n) - es(n) - z(n)), so i(n+1) = (1+ATs)*i(n) + BTs(vs(n) - es(n) - z(n) )

Here I can get F = (1+ATs), G = BTs. This is not the same with the documentation. Can you please help to explain this? Thanks...

over 7 years ago

0 Ramesh Ramamoorthy over 7 years ago

TI__Mastermind 27240 points

Your way is another way of resolving the differential equation. Either one will do the job.

0 Jack Tan80 over 7 years ago in reply to Ramesh Ramamoorthy

Mastermind 7160 points

Hi Ramesh,

Thanks for your reply, regarding to question 2, can you please help on how does F and G be gotten in DMC MATH_eSMOlib_full.pdf?

Can you please help on question 1?

0 Ramesh Ramamoorthy over 7 years ago in reply to Jack Tan80

TI__Mastermind 27240 points

If you review the transient analysis of RL circuits, you will notice that

I(t) = (V/R)*(1-exp(-R.t/L)) + I(0)*exp(-R.t/L)

Here the first coeff is G and the second is F.

0 Jack Tan80 over 7 years ago in reply to Ramesh Ramamoorthy

Mastermind 7160 points

Hi Ramesh,

Thanks for your reply.

From your equation, I could understand the G parameter, but for I(0)*exp(-R.t/L), I(0) is the initial current, in discrete form, i(n) is the last estimated current instead of initial current. So how to understand F?

BTW, Can you please help me understand what the advantage of using this transient are? What is risk of using s -> z transition, like feed forward or bilinear transformation?

0 Ramesh Ramamoorthy over 7 years ago in reply to Jack Tan80

TI__Mastermind 27240 points

Bilinear transformation fits pretty well. What is used by TI is an approximation of the time domain transient. It is a legacy carrying over for quite some time. Btw, it is one of the many ways and need not be the best. Use your judgment on what fits well for your case.

0 Ramesh Ramamoorthy over 7 years ago in reply to Ramesh Ramamoorthy

TI__Mastermind 27240 points

If your query is resolved, pls click resolved and close the thread. Thanks.

0 Jack Tan80 over 7 years ago in reply to Ramesh Ramamoorthy

Mastermind 7160 points

Thanks for your support, I've just clicked on “this resolved my issue”.

C2000™︎ microcontrollers

C2000 microcontrollers forum

TMS320F28034: Question about eSMO in DMC MATH_eSMOlib_full.pdf