3
\$\begingroup\$

I have a switch circuit which works with two phases. I need a flyback SMPS power supply for my logic circuit. I want my control circuit to be alive if any or both phases are receiving power. This can be achieved if my bulk DC capacitor receives power from both phases simultaneously. Currently, my power circuit receives power from a single phase like this:

1 phase rectifier

To add the second phase, I am planning to do this:2 phase rectifier

Will it work? I think it should because the capacitor's voltage will block any voltage lower than it's voltage right behind the diode itself, at any point of time.

Or, am I wrong? At the bulk capacitor, will there be two voltages fighting each other?

In case it works, can I do something like this:2 phase rectifier

I have removed the redundant neutral supply in the bottom rectifier, thereby saving me two diodes.

Besides this, is there anything else that I am missing out? Will the voltage on capacitor remain same whether single or both phases are supplying power or will it differ?

EDIT:

After looking at the answers, I realize that if I use full bridge rectifiers as I shown in previous images, peak output on bulk capacitor will increase significantly depending upon the phase difference of the voltages. This increase will create difficulty due to these two reasons:

1) Higher voltage means a higher voltage rated capacitor

2) It might get difficult to find a flyback IC which operates at this high voltage.

3) Even if capacitor and IC are available, it might be costlier as compared to what I can use with a lower voltage which I get in 1 phase design.

As such, I can think of another option. Using half bridge rectifiers to rectify each phase and use Neutral line as my reference:Half bridge 2 phase

Since my output power requirement is less than 5 watts (1-2 watts typical), I feel that half bridge will work even when only a single phase is available. This also ensures that my bulk capacitor will not see a voltage higher than sqrt(2)*220 V = 311 V peak. In worst case where Vrms = 270V, it will give 382 V peak. Thus I don't need to modify my existing single phase power supply circuit after the bulk capacitor.

Am I right in thinking so? Is there any better way to do this?

\$\endgroup\$
2
  • 1
    \$\begingroup\$ What is the phase angle between the two phases? \$\endgroup\$
    – Bart
    Jun 22, 2017 at 8:06
  • \$\begingroup\$ @bart - 120 degrees typically. But does it matter? In some cases, it might be difficult to tell. Example - if one phase is being supplied from UPS and other phase from grid. \$\endgroup\$ Jun 22, 2017 at 8:11

4 Answers 4

9
\$\begingroup\$

You're almost there, but not quite thinking about the problem enough in a generalized way.

If you have N leads with arbitrary AC voltage between them, then you need 2N diodes to make a full wave rectifier. Each lead is connected to one diode going to the positive DC output, and one diode coming from the negative DC output.

Note that your bottom circuit is like that, only that you've drawn it to make it not immediately obvious.

Here is a generalized full wave rectifier for N AC inputs:

You have 3 AC inputs. Your mental block seems to be thinking about the neutral line as special in that regard. It's no worse than just another AC input line, so you can make a full wave rectifier with 3 pairs of diodes.

Rectifying two of three AC phases and neutral is a odd thing to do, but is no special case for the general circuit above. Here is a plot of the voltages:

This example uses 120 V AC at 60 Hz and assumes ideal diodes.

The rectifier output is the greatest difference between any two AC inputs at any one instance. To make this easier to see, the max of all the AC inputs is shown in red, and the min in blue. The resulting difference is the green trace, which is the instantaneous output of the rectifier. This is the voltage the cap smoothes out, and is what you'd get if there was no cap and you put a small resistive load on the output.

Note that the green trace is NOT relative to the same reference as the AC inputs. It is showing the DC output when that is taken in isolation. The individual DC outputs with respect to the same reference as the AC voltages are the red and blue traces.

For contrast, here is what full 3 phase rectification looks like:

Note that this has much less ripple, and that ripple is at 6x of the AC frequency. Your two out of three phase rectification results in ripple at only 2x the AC frequency.

\$\endgroup\$
3
  • \$\begingroup\$ Whoever downvoted this, please explain what you think is wrong. \$\endgroup\$ Jun 22, 2017 at 18:15
  • \$\begingroup\$ Never mind Olin. There are all sorts of drugs out there. You never know who is high on what. Regarding my problem, I think going for a half bridge rectifier with Neutral line as reference is safest bet. That will ensure that my capacitor voltage doesn't go above 220*1.414 = 311 V (382 V at 270 VAC). Most of flyback ICs available out there will support this voltage and I can easily throw in a 450 V rated cap to handle the voltage. Power supply output power will be less than 5 watts. Am I thinking in right direction? \$\endgroup\$ Jun 23, 2017 at 5:36
  • \$\begingroup\$ @Whis: That should work. \$\endgroup\$ Jun 23, 2017 at 10:41
2
\$\begingroup\$

With your 2-phase + N rectifier scheme there will be a difference in rectifier output when running one-phase or two phases.

enter image description here

Figure 1. The voltage available to charge the capacitor will be the greatest difference between inputs. This will be much higher when two phases are present. In the 2-phase mode the neutral will carry current in the shaded zones - but even then only if the capacitor voltage has dropped below the L-N level.

With either L1 or L2 present + N you will get the corresponding voltages shown in the L to N section of the curve. This will peak at \$ V_{RMS}\sqrt 2 \$.

When two phases are present it is clear from the L1 to L2 voltages that a much higher potential difference exists and your capacitor voltage will be much higher. I will leave it to the reader to calculate the peak voltage difference between two phases.

Will the voltage on capacitor remain same whether single or both phases are supplying power or will it differ?

It will be higher with two phases.

\$\endgroup\$
6
  • \$\begingroup\$ Thanks a lot. Transistor. I think going for a half bridge rectifier with Neutral line as reference is safest bet. That will ensure that my capacitor voltage doesn't go above 220*1.414 = 311 V (382 V at 270 VAC). Most of flyback ICs available out there will support this voltage and I can easily throw in a 450 V rated cap to handle the voltage. Power supply output power will be less than 5 watts. Am I thinking in right direction? \$\endgroup\$ Jun 23, 2017 at 5:39
  • \$\begingroup\$ Your understanding is correct, I think. Don't forget that with half-bridge you only get half-wave rectification so your capacitor has to hold up longer. At 5 W this shouldn't be too difficult but in general this is not recommended as you are now pulling DC from the mains. This will generate even current harmonics on the line. See my answer to electronics.stackexchange.com/questions/312478/…. It addresses the harmonics generation for a full-wave rectifier but you should be able to figure out the half-wave situation. \$\endgroup\$
    – Transistor
    Jun 23, 2017 at 6:01
  • \$\begingroup\$ The only difference between half bridge and full bridge that I see is presence of extra diodes between mains line and capacitor. However power is always being supplied from the mains be it half bridge or full bridge. As such, I am not able to understand why half bridge will create current harmonics while full bridge will not. (Will harmonics get blocked by diodes in full bridge type?) Or do you mean that half bridge will create more harmonics as compared to full bridge? Keeping all of this aside, is there any other better way (than half bridge) to combine phases for my power supply input? \$\endgroup\$ Jun 23, 2017 at 6:59
  • \$\begingroup\$ Sorry, but length of comments is limited. Any rectification will cause harmonics as explained in the comment link. They will be worse when only one phase is connected. Update your question with a schematic of what you mean by half-bridge. \$\endgroup\$
    – Transistor
    Jun 23, 2017 at 7:04
  • \$\begingroup\$ Added the schematics and my reasoning. Please take a look and advise. Thanks \$\endgroup\$ Jun 23, 2017 at 7:25
1
\$\begingroup\$

"I think", the last schematic should work as long as the neutral is the common line.

Alternatively, you can use half wave rectification as your power demand seems low. Connect neutral to -ve terminal of the capacitor and connect each phase via diode to +ve terminal of the capacitor. Cathodes of diodes connect to +ve terminal of the capacitor.

\$\endgroup\$
1
\$\begingroup\$

The basic idea will work but there are a couple of things to be aware of.

  1. The voltage will be significantly higher when both phases are present than when only one phase is present. If the two phases are 120 degrees out of phase then the voltage will be \$\sqrt3\$ times the voltage with only one phase present. If the two phases are 180 degrees out of phase then the voltage will be twice the voltage with only one phase present. Your capacitor will have to be selected (the one you have selected has far too low a voltage rating) and your fly back converter designed with this voltage range in mind.
  2. If the neutral becomes disconnected or one of the diodes fails then there is a potential for back-feeding. This may or may not be a safety hazard depending on the larger context of the system.

Overall I suspect you are probably better off using two separate isolated power supplies and putting your combiner-diodes on the output side.

\$\endgroup\$
2
  • \$\begingroup\$ Thanks Peter Griffin. I am thinking of using Neutral line as reference and put half bridge rectifier on both the incoming phases. This will give Peak output of 220*1.414 = 311 V on the capacitor. 382 V at 270 VAC (worst case scenario). The voltage will remain in this region irrespective of their phase difference. Am I thinking right? \$\endgroup\$ Jun 23, 2017 at 5:48
  • \$\begingroup\$ Yes, peak is not affected by phase assuming negligible load. \$\endgroup\$
    – nvd
    Jul 5, 2017 at 21:16

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.