1
\$\begingroup\$

So I want to build a cheap AC-DC buck converter and I'm trying to eliminate parts by replacing them with "smart switching". So my idea is to ignore the capacitor which would smooth out the rectified AC and instead have the high frequency switching device be aware of the AC signal and switch on and off accordingly to smooth out the signal.

This works fine as long as there's power to draw from, however during the dips in the rectified AC signal there's no power to draw from at all so my question is if there's any "easy" (cheap) way to phase shift the power by 90 degrees so that there's always power for the high frequency switcher to draw from? Or is there some other alternative than adding a large capacitor to smooth out the signal?

I've tried to look around for phase shifting power, but all I get are the expected "single phase - 3 phase" stuff.

\$\endgroup\$
6
  • 1
    \$\begingroup\$ That sounds just like normal PFC with extra steps. \$\endgroup\$
    – Jeroen3
    Commented Feb 27, 2020 at 13:56
  • 5
    \$\begingroup\$ Phase shifting probably involves a large capacitor. Just saying..... \$\endgroup\$ Commented Feb 27, 2020 at 13:59
  • \$\begingroup\$ And you'd have to shift it by 180 degrees to provide power during the "off" portion of a simple rectifier. \$\endgroup\$
    – rdtsc
    Commented Feb 27, 2020 at 14:03
  • \$\begingroup\$ @rdtsc 90 degrees of the input AC signal, but yes 180 for the rectified signal. \$\endgroup\$ Commented Feb 27, 2020 at 14:07
  • 1
    \$\begingroup\$ The easiest way (without an inductor or capacitor) is to start with a 3-phase supply. \$\endgroup\$
    – user16324
    Commented Feb 27, 2020 at 14:30

2 Answers 2

2
\$\begingroup\$

** Or is there some other alternative than adding a large capacitor to smooth out the signal**

Adding a large filter capacitor at the input is largely avoided in modern designs, so you are not trying to do something very different to normal practice in that respect.

You may benefit by looking at the implications of "smooth output".
If you are trying to maintain power out as Vin falls then you need to draw increasing current as Vin sinusoidal falls. You get a sinusoidal offset current peaking to very large values as Vin approaches zero.

The alternative to supplying phase shifted AC is to store DC energy - and a suitably sized reservoir output and maybe input capacitor plus perhaps additional LC filtering is the normal method - not without reason. As Andy notes - drawing increasing current as Vin falls is the opposite of good power factor compensation practice and would get your device rapidly noticed and explicitly contrary to modern regulatory requirements.

That said, you may wish to look at valley fill circuits which attempt to do something like what you describe but on the DC side of the converter.

Somewhat more focused subset of the above.

And similar

Simple valley fill circuit to maintain DC level at converter imput as AC voltage falls:

enter image description here

Getting fancier

"HV9861ADB2, HV9861A LED Driver Demo Board Boost Assisted, Valley Fill 120VAC Input, 7W Output, 350mA, 20V, Power Factor 93% Reference Design using part HV9861ALG-G by Microchip Technology" - From here

enter image description here

This paper is relevant. Paper is present if you scroll down.

\$\endgroup\$
0
\$\begingroup\$

I want to build a cheap AC-DC buck converter

Restricting your self to using a buck converter (as opposed to a buck-boost converter) does mean that as soon as the rectified AC voltage waveform falls close to the DC output of the buck circuit, no energy can be transferred.

But, if you used a buck-boost regulator, the rectified AC voltage waveform could drop lower and there would be still energy transferred to the DC output. Not all the way to zero volts of course but, a significant improvement especially when the DC output is tens of volts as opposed to single digit volts.

However, this is contrary to modern philosophy where an important consideration is to design power supplies that take a current waveform from the AC that looks like a resistive load is attached (aka power factor correction).

any "easy" (cheap) way to phase shift the power by 90 degrees so that there's always power for the high frequency switcher to draw from?

Again, this is contrary to modern "green" thinking so, you have to decide what your priorities are.

\$\endgroup\$
1
  • \$\begingroup\$ See my valley fill comments - sort of what he is trying to do on the load side. Sort of. ... \$\endgroup\$
    – Russell McMahon
    Commented Feb 27, 2020 at 14:51

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.