1
\$\begingroup\$

Im trying to develop the PWM-Switch Model of a Cascaded-Buck-Boost.

First of all i want to extract the Large-Signal Transfer function for the Buck-Mode.

I'm working through the book "Switch-Mode Power Supplies_ SPICE Simulations and Practical Designs" from C. Basso (here: https://cbasso.pagesperso-orange.fr/Spice.htm)

and the APEC 2013 Seminar: https://cbasso.pagesperso-orange.fr/Downloads/PPTs/Chris%20Basso%20APEC%20seminar%202013.pdf

My problem is, i cant get the jump of the Topology from Slide 57 (in the Seminar) to Slide 61 (Maybe i think Slide 57 is the Buck, and 61 is Boost?)

But anyway, I have problem rearraging my Topology to see clearly what my next step is .. maybe someone can help me to get further:

Whiteboard

\$\endgroup\$
3
  • \$\begingroup\$ I think the author of that book sometimes frequents this site. It might be @verbal kint? \$\endgroup\$
    – Andy aka
    Commented Oct 31, 2019 at 17:59
  • \$\begingroup\$ "Maybe i think Slide 57 is the Buck, and 61 is Boost?" - slide 57 shows a buck converter, but slide 62 is 'any 2-switch converter', and slide 62 shows a boost converter. Difference is buck converter has switch before inductor, boost converter has inductor before switch. \$\endgroup\$ Commented Oct 31, 2019 at 21:13
  • \$\begingroup\$ @BruceAbbott: okay, I already thought this would be the reason, but it was a little bit confusing. But thanks for the reply! My problem now is, that I don’t know how I can rearrange the circuit to get the large-signal Model transfer function \$\endgroup\$
    – rst_tk
    Commented Oct 31, 2019 at 21:15

2 Answers 2

2
\$\begingroup\$

By looking at this cascaded buck-boost, I believe two PWM switch models are needed and must be assembled as below:

enter image description here

To verify if the averaged models arrangement is correct in large-signal, I usually plot the step-load response or the input response and compare it to that of a simple open-loop-operated cycle-by-cycle model. If the average model is correctly wired then responses must be very close to each other. Please not that I did not include the PW modulator as the ramp amplitude is assumed to be 1 V hence a gain of 1.

Here, I can use the SIMPLIS demo version elements and run the complete circuit then obtain its small-signal response. The results are shown below and agree well in the operating point. The voltage is 12 V for a nearly 50% duty ratio.

enter image description here

Regarding the ac response, it is very close to what the average models give:

enter image description here

What is the next step then? As these are large-signal models, you need to replace both PWM switch models by their small-signal version and rearrange the whole circuit in an easy-to-read form. I would recommend to use a simulator first (LTspice can do well) and store the first ac response obtained with the large-signal models: that is your reference. Then replace the large-signal models by the small-signal ones. Then simplify, rearrange your circuit to make it easy for you to analyze. Each step that you take when you simplify or rearrange the circuit must be backed up by a new simulation that you compare with the reference. Any significant deviation will highlight either a mistake when rearranging the circuit or an inappropriate simplification (a parasitic term you decide to ignore etc.). To that respect, simulators are an incredible asset when doing these exercises.

Once you have a clean circuit simple and well readable, then you can start looking at its control-to-output transfer function or any other. To do so, I would strongly recommend you learn how to apply the fast analytical circuits techniques or FACTs. Simply put, you cannot beat them at execution speed and ease of implementation. It requires a bit of time and practice to acquire the skill but when you have, there is no going back. And for the DCM version? Simply replace the PWM switch models by their DCM small-signal versions and you restart the analysis. As Marko would say, "good luck"!

\$\endgroup\$
0
0
\$\begingroup\$

thanks for your effort! I have some Questions about your way to go:

First of all, you have inserted 2-PWM-Switchs -OK! enter image description here

But isnt the Diode there in the wrong way? In the Block PWM Switch VM the Diode represents my Diode D1. The Second PWM1 Switch VM-Block shows the my Diode D2 in the wrong turn. Or what im missing?

The goal will be, if the incoming Voltage is lower than my reference Voltage, the Boost-Controller will take action (Buck-Switch will always closed). If the incoming Voltage is too high, the Buck-Controller will take action (Boost-Switch always open). So i thought I can extract the Large-Signal TF and after that the Small-Signal TF for the Buck-Modus first (see my Circuit, Sboost is open, Diode D2 is conducting). And after that the same procedure for the Boost-Modus.

I gather you would insert 2 PWM-Switches simultaneously and solve for it. But for me I think it is more complicated solving 2 PWM-Switches, instead of 1 for every Mode, or what do you think?

\$\endgroup\$
2
  • 1
    \$\begingroup\$ Using an answer to ask questions about another answer is not the correct way to go. Please use the comment section below mine to ask about the implementation. Also, if you consider my answer as appropriate, please credit it accordingly. There is nothing wrong in the wiring of the 2nd PWM switch: the diode is just a symbol and current can flow in any direction. If both switches are actuated together, as I did, you cannot separately solve the blocks and combine individual responses, you need to work with both small-signal models in place and determine the corresponding transfer function. \$\endgroup\$ Commented Nov 4, 2019 at 12:35
  • \$\begingroup\$ Sorry, you are Right. I have too less reputation to +1 your answer. Ok, thanks for the Info! That’s good to know, I will go ahead! Thanks for your help! \$\endgroup\$
    – rst_tk
    Commented Nov 4, 2019 at 12:39

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.