0
\$\begingroup\$

I'm still working on designing my push-pull converter, for which I’ve already managed to determine the open-loop transfer function in this question: How can I measure the transfer function of a push-pull converter in SIMPLIS?

Now I want to close the loop using a non-inverting type 2 compensator. I prefer the non-inverting configuration to avoid the fast-lane effect.

enter image description here

I chose a crossover frequency of 2 kHz and a phase margin of 65 degrees. The transfer function of the complete compensation network is:

$$H(s)=G_0 \cdot G_1 \cdot \frac{1+ \frac{s_z}{s}}{1+ \frac{s}{s_{p1}}}$$

Where \$G_0 = \frac{R_6}{R_7} \cdot \frac{R_{11}}{R_{11}+R_{10}} \cdot CTR\$

and \$G_1\ = \frac{s_{p0}}{s_z}\$

I determined the zeros and poles using the "k factor." At the crossover frequency, I had to set a -20 dB attenuation, which I achieved by adjusting \$G_0\$ and \$G1\$. This resulted in the value of \$f_{p0}\$ being \$1.937 kHz\$. The additional pole and zero are: \$f_z = 0.748 kHz\$

\$f_{p1} = 5.349 kHz\$

However, the POP simulation still doesn't run because it can't converge. What could be the issue? How can I simulate just the feedback loop without the other circuit blocks?

EDIT:

I managed to simulate the error amplifier. The gain and phase margin matched the calculated results, so that part is definitely correct.

enter image description here

I ran a transient simulation as VerbalKint requested, and it seems as if the error amplifier isn't working at all. Vfb is controlling the modulator with a constant 5V.

enter image description here

enter image description here

EDIT2:

I replaced the non-inverting Type II error amplifier with an inverting one. Unfortunately, the simulation still doesn't work, even though the error amplifier is configured correctly. Unfortunately, I don't know how to proceed. :(

enter image description here

enter image description here

enter image description here

\$\endgroup\$
2
  • 3
    \$\begingroup\$ Hi, I would first try to run a transient simulation to see if the output at least rises towards the regulation level and see if the loop takes over at some point. If not or if the voltage does not move in the right direction, there is no way the POP algorithm can converge. Regarding the fast lane, you know that you can disable it easily but you can also look at the structure adopted in this NCP1566 demo doc. \$\endgroup\$ Commented Jun 27 at 16:57
  • \$\begingroup\$ @VerbalKint I ran the simulation and attached the images to the question. I also managed to simulate the error amplifier, and it works as expected. \$\endgroup\$ Commented Jun 27 at 19:14

0

Reset to default

Your Answer

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

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.