2
\$\begingroup\$

Context: tuning the control loop of a buck SMPS. I specicifally have a TI TPS40200-based circuit in mind, but I think this applies to SMPS design in general.

Quoting from Type 2 compensation network:

Mr. Maniktala suggests that the loop gain transfer function of the switcher intersect the 0 dB point at approximately 1/6th the switching frequency, with a slope of -1.

I have heard this before from another engineer, but did not get a chance to ask why. What is the reason the crossover frequency of the control network should be a fraction of the switching frequency? Why specifically around 1/5th or 1/6th?

Are there any other desirable traits when designing a compensation network for an SMPS controller other than the most basic criterion of >45 degree phase margin for system stability?

\$\endgroup\$

1 Answer 1

Reset to default
5
\$\begingroup\$

A switch mode power supply is essentially a sampled-data system, therefore the theoretical maximum bandwidth is one half the switching frequency. Practically the phase and transport lag there make it impossible to close the loop there, so 1/5 to 1/10th the switching frequency is a good rule of thumb.

There are many other considerations in compensating an SMPS- Gain margin, conditional stability, current vs. voltage mode, slope compensation, transient response, etc.

Check out www.ridleyengineering.com, there are lots of good free tutorials and papers there.

\$\endgroup\$
2
  • 3
    \$\begingroup\$ Ah. So I was overlooking the fact that a sampled-data system need not be a digital control system. You can only update a PWM duty cycle once per period, and that makes this a sampled-data system because you output is sampled, even if the controller itself is almost completely analog circuitry. This is a case where the terms 'discrete-time' and 'digital' are far from interchangeable. \$\endgroup\$
    – Dmitri
    Commented Aug 26, 2014 at 19:27
  • \$\begingroup\$ Exactly! You can use a zero-order hold transfer function in the z domain as part of the control loop model of a SMPS to model the sampling effects. \$\endgroup\$
    – John D
    Commented Aug 27, 2014 at 3:05

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.