I'm working in a buck converter to step down 127VDC to 5VDC/1A. I know that boost converters have limitations, thus it's not recommended to step up 5VDC to 127VDC. But I couldn't find anything about buck converter limitations. After Reading this document from TI I figured out some values for inductors, capacitors and so on. But my doubts still remain: Is it okay to step down that amount of volts? Will I find any trouble with my common buck converter circuit?

Obs.: The load will be constant and it is not sensitive to voltage oscillation around the 5V tension.


simulate this circuit – Schematic created using CircuitLab

  • 2
    \$\begingroup\$ A usb power supply would probably work just as well. \$\endgroup\$ Oct 1, 2017 at 0:11
  • \$\begingroup\$ Yeah, I know. But I really would like to learn more about buck converters and it's possible applications on rectified mains voltage. So, no easy way this time :3 . \$\endgroup\$
    – salgado
    Oct 1, 2017 at 0:18
  • \$\begingroup\$ High ratio can be achieved but is not common due to the limitations several people have mentioned. I designed a 12-200 VDC in / 12 DCout buck converter for use in exercise equipment that used an alternator as the user load. It worked but efficiency was low near Vout max. Still FAR better than a linear regulator. \$\endgroup\$
    – Russell McMahon
    Oct 1, 2017 at 11:35

1 Answer 1


One problem with a large voltage ratio buck converter is that the duty-cycle becomes very small - the ratio of the voltages - and if you maintain a reasonable frequency of operation the ON time gets very short.

Also the way you have it here with an N-channel FET the PWM signal is required to be over 130V in amplitude - this will be difficult to create, especially as it is above the maximum Vgs of the switching FET.

A buck converter is non-isolating so the output is galvanically connected to the AC input - potentially a safety issue.

Also - why do you specify 127V as the input? If it is really AC line power at 120V when you rectify it you will get ~165V.

The normal solution to all these problems for low to medium power is a flyback converter using a transformer. These can be very simple, although it is usually simpler to purchase a ready made unit.

Simple line powered flyback SMPS

From SMPS schematic

Design Guide for Flyback Converter

  • \$\begingroup\$ Yeah, you are right man, I forgot to edit that voltage values (in fact, AC lines are 127V here). According to that TI paper that I mentioned, the duty cicyle will be 3.3% and frequency 100kHz. Is it possible to work? Considering the modifications that you proposed for the MOSFET. I will consider the flyback design, I really liked the idea. Thanks for the answer! \$\endgroup\$
    – salgado
    Oct 1, 2017 at 0:27
  • 1
    \$\begingroup\$ Another drawback of high voltage ratio buck (and boost) converters is the need for both switching devices (MOSFET and diode in this case) to withstand both the full input voltage and the peak inductor current. In a transformer -based DC/DC converter (like a flyback) this isn't the case. \$\endgroup\$
    – jms
    Oct 1, 2017 at 0:30
  • \$\begingroup\$ @Kevin White, not sure about the winding dots but considering the two diodes and the output inductor \$L_2\$ it looks like you've inserted the picture of a forward converter : ) Most of the small non-isolated converters found in the white-goods market are buck converters (up to 276 V rms). A tapped version can help for the duty ratio but the flyback converter is the easiest way for low-power isolated stuff. \$\endgroup\$ Oct 1, 2017 at 7:51
  • \$\begingroup\$ @VerbalKint - good catch. My error - Sorry about that, more haste less speed. \$\endgroup\$ Oct 1, 2017 at 17:09
  • \$\begingroup\$ @Kevin White, no problem. I can see you have inserted the pix of a nice ringing-choke converter. I remember seeing these in cell-phone chargers back in 1997. They don't like to be unloaded (at least this simple version) hence the fairly chubby bleeder in the output. \$\endgroup\$ Oct 1, 2017 at 18:22

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.