1
\$\begingroup\$

How do I implement the zero current detectors (ZCD) in the buck converter?

I want to use it to implement the ZVS buck converter, and I need to detect the zero current, but I have no idea about ZCD.

Can someone give me some idea?

Here is a ZVS buck converter article: Control and Analysis of Synchronous Rectifier Buck Converter for ZVS in Light Load Condition

enter image description here

\$\endgroup\$
4
  • 2
    \$\begingroup\$ The ZC happens only in DCM mode, and the switches have already switched their states. In no way you could achieve a ZCS condition with this buck. \$\endgroup\$ May 18, 2021 at 9:31
  • 2
    \$\begingroup\$ In the answer I gave you here there is a SIMPLIS simulation schematic. If you want to detect a zero current condition in the inductor - for whatever usage - you just add an extra winding on that inductor and you can easily detect when the current hits 0 A. Other methods are possible with a current sense resistor but it requires negative sensing and complicates a bit the electronic circuit. \$\endgroup\$ May 18, 2021 at 9:58
  • \$\begingroup\$ @VerbalKint Thanks for your answer. I know how to do \$\endgroup\$
    – Jitter456
    May 18, 2021 at 11:37
  • \$\begingroup\$ @VerbalKint I update the article it talks about ZVS buck converter, and it said the optimal delay time is 1/4 of the resonance cycle. do you know why? \$\endgroup\$
    – Jitter456
    May 18, 2021 at 11:45

1 Answer 1

3
\$\begingroup\$

Quasi-square-wave converters also known as quasi-resonant (QR) converters are well known to offer zero-voltage switching conditions or ZVS to the power switch in specific conditions.

A borderline-operated buck converter can be operated this way in a self-relaxing system where the inductor core reset is detected via an auxiliary winding wound over the main power inductor. A few turns are enough to observe a voltage induced by the flux swing inside the core: \$v(t)=-N\frac{d\phi(t)}{dt}\$. When the core is demagnetized (or saturated by the way), there is no more flux variations and the voltage on the winding should be zero volt.

In reality, because of the parasitic capacitance, the voltage across the power switch cannot instantaneously return to \$V_{out}\$ in a buck converter and an energy exchange takes place with the circuit inductance and capacitance. Deacaying oscillations made of valley and peaks occur across the power switch. When it goes through a minimum - a so-called valley hence the term valley-switching operation it is there where turn-on losses would be at a minimum. These oscillations are observed across the auxiliary winding centered at 0 V and when the voltage crosses a particular threshold, the MOSFET is turned back on. Below is a typical waveform across a flyback converter operated in QR (excerpted from here):

enter image description here

You now need to calibrate the internal 0-V detector (usually it is a fixed 40-60-mV threshold) so that it turns the MOSFET exactly in the valley. A small delay is thus added so that the threshold on the auxiliary winding corresponds to the valley in the switch. If you look at the below simulation, you see this delay counted when the voltage approaches 0 V on the auxiliary winding is 1/4th of the oscillation period if you want to switch in the 1st valley:

enter image description here

This is a self-relaxing borderline-conduction mode (BCM) buck converter as described in the free 60+ ready-to-use SIMPLIS platforms described here. A paper describing how to operate a buck in ZVS can be found here.

\$\endgroup\$

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.