Loop Gain Meter weird behaviour

General
1

Hello all,

I am using PLECS version 4.6.9 and I am observing a weird response from the Loop Gain Meter block.
I want to evaluate the closed loop frequency response and there is a significante difference between using this block or using the separate Small Signal Pertubation and Small Signal Response blocks.
I am comparing three cases using the template model ‘buck_converter_with_loop_gain_analysis’ and beside the orignal block, in the loop, I’ve added two others Pertubation / Response pairs:

image
image773×393 12.2 KB

The response are:

The model is:
buck_converter_with_loop_gain_analysis.plecs (33.2 KB)

Can anyone help sort out the difference from the Pertubation / Response results from the Loop Gain Meter?

Thanks.

2

The open-loop gain measures something fundamentally different than other gain meters. In the first graph, the loop is open and you’re measuring G(s). In the second two graphs you’re measuring 1/(1+G(s)).

You can look under the mask of the loop gain meter to see its implementation. Also from the help section of the Response block, which determines Y(s)/U(s): “If the reference input is shown, U(s) is calculated from the signal that is connected to it. Otherwise, U(s) is calculated from the perturbation signal generated by the corresponding Small Signal Perturbation block.”

  • The loop gain meter “breaks” the feedback path by using the reference input of the Response block. The response block input U(s) is effectively “-e” and then Loop block in the figure calculates vo’/e=G(s) or the open-loop gain. Note the negative sign on the “e” signal is dropped as the response block has the reference input inverted to account for the sign of the summation node.
  • The second graph shows the closed loop gain or vo’/~ref = 1/(1+G(s)).
    • vo’ = e*G(s)
    • e = (~ref - vo’)
    • vo/~ref = 1/(1+G(s))
  • The third response is essentially the same as the second. One can show e=(ref + ~ref - vo) is equivalent to e=(ref - (vo + ~fb)) when you account for the sign difference (which you did with the -1 gain infront of Vout_I) and ~ref and ~fb are the perturbation input defined in the analysis tools settings.

image
image348×144 2.78 KB

3

Thank you for confirming.

I now understand that it is a open loop gain measurement and not a close loop gain as we may understand from the Help: “Small Signal Gain: Measure loop gain of closed control loop using small-signal analysis”