buck boost converter and others

[sarahkarigan]

For switch mode power supply, there are buck-boost converter, sepic converter, flyback converter and cuk converter, they all can buck and boost an input. how does one determine which converter to be used in his or her design? thank you

 

[buckysparks]

most develop a specification for the power supply from the requirements: input voltage range, output voltage, output current, output voltage ripple, efficiency goal, physical size constraints , costs, manufacturing limitations ( eg thru hole, smt, leadless components) etc
based on these one can determine the better topology

 

[sarahkarigan]

buckysparks, thanks for your reply. how do i know which topology is most efficient? thx

 

[Claude Abraham]

Buck-boost - 2 FETs, 2 diodes, 1 or 2 inductors, single stage energy conversion, 1 complex conjugate pole pair in ccm (continuous conduction mode operation), 1 right half plane zero in ccm, moderate bandwidth, noisy input & output require pre- & post-filtering, low efficiency at low output voltage due to 2 diodes, moderate efficiency at higher output voltages.

Cuk/boost-buck - 2 FETs, 2 diodes, 2 inductors, dual stage energy conversion, 2 complex conjugate pole pairs in ccm , 2 right half plane zeros in ccm, low bandwidth, quiet input & output require no pre- or post-filtering, low efficiency at low output voltage due to 2 diodes, low to moderate efficiency at higher output voltages.

Sepic - 1 FET, 1 diode, 2 inductors, dual stage energy conversion, 2 complex conjugate pole pairs in ccm, 3 right half plane zeros in ccm, lowest bandwidth, quiet input & noisy output require no pre-filtering but requires post-filtering, moderate efficiency at low output voltage due to 1 diode, moderate efficiency at higher output voltages.

Isolated flyback - 1 FET, 1 diode, 1 transformer gapped to store energy, single stage energy conversion, 1 complex conjugate pole pair in ccm, 1 right half plane zero in ccm, moderate bandwidth, noisy input & output require pre- & post-filtering, moderate efficiency at low output voltage due to 1 diode, moderate efficiency at higher output voltages. Similar to buck-boost except isolated due to xfmr.

**broken link removed**

 

[sarahkarigan]

Claude Abraham, thanks for your reply. How do know about the efficiency performance at different output level? Is there any article or sources that I can refer to? thanks again!! =)

 

[Claude Abraham]

I haven't written it yet, but I might have time soon. In a nutshell, at low output voltages the sepic has the lowest overall losses due to just 1 FET & 1 diode. But the FET & diode must withstand the full output voltage plus maximum input voltage. A buck-boost has 2 FETs & 2 diodes. The 1st FET & diode must withstand the maximum input voltage, The 2nd FET & diode only need withstand the output voltage plus a diode drop (for the FET). The boost-buck has 2 FETs & 2 diodes, all of which must withstand the full output plus maximum input voltage. The isolated flyback has 1 FET & 1 diode each of which must withstand double the input voltage or even more depending on duty cycle.

In a nutshell, sepic losses are lowest at low outputs, buck-boost is a tad better at moderate to high outputs, boost-buck/Cuk has greateer losses as well as isolated flyback. Still, all can be pretty good in terms of efficiency. Boost-buck/Cuk is the quietist but slow, buck-boost & flyback are the noisiest & moderate speed, with sepic being quiet at input noisy output, and slowest speed of all. I hope this helps. I've edited previous post adding references I published. BR.

 

[sarahkarigan]

Thanks Claude. I have few more questions:
1) What defines high output voltage, low output voltage that you mentioned? is it relative to the input? i.e. output is higher or lower than input? or is there a range that categorize them like stated in wiki: Low voltage - Wikipedia, the free encyclopedia

2) What does it mean that the converter has slow speed at the output? does it mean the transition from buck to boost or vice versa is slow?

3) Since SEPIC has the slowest speed of all, is it not suitable for thyristor gate trigger circuit design?

thanks =)

 

[Claude Abraham]

1) "Low" output voltage is one where the forward voltage drop of the diode(s) is significant, i.e. 5V, 3.3V, etc. At 12V, the difference between 1 & 2 Schottky diodes is less important than at 3.3V.

2) Slow speed refers to the servo loop 0-dB bandwidth. Should the load current demanded by the output suddenly change, there is a time required for the loop to restore the regulated value of output voltage.

3) It depends on the switching frequency. If the freq is low, like 1.0 kHz or less, then a sepic should work if designed properly. If the switching freq is 400 kHz, then no. The exact limit can only be determined by computation. A sepic which delivers low output current, i.e. less than 1.0 amp, may be designed for discontinuous (dcm) operation. This eliminates 1 complex conjugate pole pair & the right half plane zeros. Bandwidth can be greatly increased since stability is less of a problem. It depends on freq & output current.