0
\$\begingroup\$

So this question is a bit related to my other question on using Pspice to get the required transfer function of the buck converter.

Here in this circuit I'm trying to get the open loop transfer function bode plot of the converter. As suggested by VerbalKint in the other question, I am using the dB(voltage) probe as shown (not using x,y points):-

enter image description here

And this is what I get :- enter image description here

The Pspice bode plot is wrong, since there should be an overshoot of nearly 32 dB at the resonant frequency 7.587 kHz, but here its all wrong. : An error I realized pretty late, I should be using the DCM buck transfer function.

Here is the circuit with bias voltage display.

enter image description here

This is what I got on MATLAB using the DCM Transfer function:- enter image description here

So it is still quite different, as from Pspice I get PM = 48.73 degrees at 89.03 kHz, very different from MATLAB.

EDIT 1:- Here is the model definition for U3 used here:- enter image description here

\$\endgroup\$
12
  • \$\begingroup\$ What is the computed dc point? Always check it first before considering any ac results. \$\endgroup\$ – Verbal Kint Mar 5 at 18:32
  • \$\begingroup\$ I am not asking for the transient results but for the dc operating point computed prior to the ac analysis with the averaged model. OrCAD offers the ability to reflect the bias points to the schematic nodes once the ac analysis is done. This is a good practice to adopt for any ac analysis. If the dc operating points are wrong, SPICE can still deliver an ac response but it is likely to be garbage. \$\endgroup\$ – Verbal Kint Mar 5 at 19:49
  • \$\begingroup\$ @SM32 Looks like trouble: the output voltage is not 200 V. Are there parameters to set in the model? \$\endgroup\$ – a concerned citizen Mar 5 at 20:34
  • \$\begingroup\$ @VerbalKint did you mean the image I added showing the voltages at different nodes? Because I enabled the "bias voltage display" button, sorry if my question is silly, I'm new to Pspice simulations.here the output is not showing 200V, but the same thing in transient analysis did show 200V,so I can't get what's going wrong. \$\endgroup\$ – SM32 Mar 6 at 5:54
  • \$\begingroup\$ @aconcernedcitizen The transient analysis circuit is giving me 200V, the values agree with the formulas too.I haven't added any parameters, I assigned them from the very beginning. I think it is going wrong in the way I'm using CCMDCM1; my actual circuit is a synchronous buck,but here it's a MOSFET & diode config.Would that be a problem? \$\endgroup\$ – SM32 Mar 6 at 5:56
2
\$\begingroup\$

I have captured your circuit using the PWM switch model in voltage mode and its equivalent differently derived by the CoPEC authors. The circuit is shown below:

enter image description here

First you need to check the dc points and confirm that they are correct and similar between the models. Both circuits deliver 270-271 V which is ok for a 50% duty ratio. We can now ac sweep both circuits and check their frequency responses:

enter image description here

The control-to-output transfer functions are rigorously identical which confirms that models are ok. I prefer symbols where you see the internal arrangements, it is easier to check for mistakes later on in my opinion.

If I now apply the DCM formula that I have derived in my next book on small-signal modeling, the Mathcad sheet delivers a response similar to that of SPICE:

enter image description here

As you may know, I am releasing a new book on small-signal modeling of switching converters (TOC is here) in which I derived many transfer functions. The one for the DCM buck operated in VM was originally derived here.

enter image description here

\$\endgroup\$
6
  • \$\begingroup\$ Thank you very much for this. I'm consistently getting the DCM transfer function wrong which is the source of my troubles, could you tell me the final transfer function you derived? I'm going wrong with the rc,rl,rdson. \$\endgroup\$ – SM32 Mar 7 at 18:28
  • \$\begingroup\$ I'm getting Phase Margin(PM) = 90.2 deg @ 82.2 kHz, which is a bit different from the Pspice results of PM = 86.338 deg @ 98.25 kHz. Unless this is acceptable, if you could please give me the final transfer function without the derivations also, it'd be great. In this TF, the far away pole hasn't been taken into account. \$\endgroup\$ – SM32 Mar 8 at 11:59
  • \$\begingroup\$ I have derived all 4 transfer functions of switching converters in my next book (see the TOC). I'll try to disclose the control-to-output TF of the DCM VM buck tonight. \$\endgroup\$ – Verbal Kint Mar 8 at 12:27
  • \$\begingroup\$ Yes,I've seen the TOC before also,it has a lot of important things! Thank you, the TF (expression in variables,i.e generalized) would be most helpful! \$\endgroup\$ – SM32 Mar 8 at 17:36
  • \$\begingroup\$ Thank you very much, but unfortunately I'm still a bit stuck. I think denominator of wp1 is 1-M. Also, if M = Vo/Vin, what should be Duty cycle D? I mean, if I take D=M=0.5, I'm getting it wrong. I tried D = pulse width * frequency, but that's giving the wrong crossover frequency. \$\endgroup\$ – SM32 Mar 11 at 18:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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