We have three phase input and the voltages in the three input is not the same always. Sometimes, one or two or three phase will be off. I would like to connect the phase to output which lies within the range (110 - 290).

Below will be the example,

Phase 1 (290v), Phase 2 (245v), Phase 3 (190v)

I need to select Phase 1 in this case. Ampere rating would be 40A on the selected output.

Did searched in google for single device (circuit / IC) which can does this, not finding one. Any direction would be helpful.

Voltage Reference: (Added)

Voltage reference is to the neutral line. It would be a

4-wire input to 2 wire output with a common neutral line.

Circuit Thoughts:

Thought 1:


Compare the voltage on each phase with a voltage comparator, connect the output with a solid state SSR in the output.

I would need a power supply now that can input from magic three phases and generate DC output. Not finding the right one.

Will that work?

Thought 2:

This seem to be a cheaper and reliable solution, without inductive circuits. Please suggest if I have missed something to consider. SMPS - Switch Mode Power Supply (5V).

Thought 3:

This is the cheapest solution, if the input voltage is constant. With the varying voltage, the coil does not seem to take a varying voltage. (100-300V). Help from https://electronics.stackexchange.com/a/71066/24465 . Thanks to the community. If you think something can be improved, please share your thoughts.

  • \$\begingroup\$ Sounds like you have a floating null. How are you measuring the voltages, between which wires? \$\endgroup\$ – jippie May 27 '13 at 19:56
  • \$\begingroup\$ do you want uninterrupted transfer from one phase to another(fast phase change over time or slow change over time) \$\endgroup\$ – yogece May 28 '13 at 20:32
  • \$\begingroup\$ Slow or fast would be fine as long as output is within the 290v instead of no power. Would prefer a SSR with no mechanical movements. \$\endgroup\$ – Kannaiyan May 29 '13 at 0:35

Conceptually this is an easy problem, the reality is that you are dealing with relatively large load currents (40 A) and several 10's of KW. This all combines to very large components, with relay like switch gear, switching between phases which will cause glitches and drop outs just through the switching action. Also, I suspect that is a very dynamic situation, and having relay contacts hammering away at the phases are dynamically changing will wear them out very quickly.

Any way you look at it this is not a trivial problem to solve.

The one solution I can think of that will provide the smoothest transition between phases is to duplicate a non-bypass type UPS (Uninterruptible Power supply). This would comprise of a 3 input full wave bridge rectifiers that are generating high voltage ripple DC, the AC input with the highest voltage automatically contributes the most. And then you run an inverter that generates a stable 220 V AC @ 40 A from that HV DC rail.

Designing and building such an inverter is not a trivial task, but perhaps you can buy such a unit.

Because of the wildly fluctuating inputs (from 110 - 290 V) it might be better to have 3 transformers @ 2:1 step down ratio, 3 bridge rectifiers that generate your HV DC rail at ~ 110 V DC ripple that then is fed to your inverter which boosts it to 220 AC. But this is even a harder design at these high power levels.

It is entirely possible that one of these stabilizer units that you linked to could possibly fill in as your inverter. But it very much depends upon how they are designed.

  • \$\begingroup\$ Given a Three phase input with 4 wire, if there is a three phase comparator can give a digital output like 001, 010, 100. I can straightaway connect it to an SSR (cheaper in ebay) and get the grid input directly to a single phase voltage stabilizer. Voltage stabilizers does not create new waveforms, it increases or decrease the amplitude of the incoming signal. There is a little bit of delay between the grid input and voltage stabilizer output. \$\endgroup\$ – Kannaiyan May 29 '13 at 3:14
  • \$\begingroup\$ Checked online for 40A pure sine wave inverter, seems too bulky and expensive. Converting three phase to DC and converting back to AC, seems an expensive task. The need which I'm discussing here is for the whole house supply. Peak could be 40A. \$\endgroup\$ – Kannaiyan May 29 '13 at 12:52
  • \$\begingroup\$ You need to make sure that your comparator circuit operates over longer phase times or you'll be attempting to switch on every phase peak but you also need to be able to switch within a phase time as a phase could possible fali within a cycle time (but hopefully not). By switching between phases you will be generating glitches and spikes and the reactive components of the stabilizer will possibly generate spikes of their own (inductive load). if it switches frequently enough you'll be generating a C component as well. All things to consider. \$\endgroup\$ – placeholder May 29 '13 at 14:02

Not sure about the English terminology, but this is what I mean:

enter image description here

In a graphical representation of a three phase system, the three phases can be viewed as the corners of a triangle. When the load on all three phases is identical, null will be exactly in the middle of the triangle with equal voltages to each phase 230V in the image above.

However when the load differs per phase, the null will shift away from the middle and the voltage measured will be different per phase. In above example, in ideal situation you measure 230V between any phase and null. In worst case the voltage between a single phase and null may rise to up to 380V.

Cause is a not properly connected null wire. As voltages may rise to 1.4 times the rated value you really want to get it fixed. Either properly connect null to the grid or maka an equal distribution of the loads over the phases.

In Dutch it is called "zwevend nulpunt", but I'm unsure about correct English terminology.

  • \$\begingroup\$ Added the voltage reference. Output does not need a phased output. It can be from any one of the input. I want to compare high voltages and provider the higher voltage at the output if that is within the output range. \$\endgroup\$ – Kannaiyan May 27 '13 at 20:16
  • \$\begingroup\$ I think you should fix the cause, not remedy the result. \$\endgroup\$ – jippie May 27 '13 at 20:18
  • \$\begingroup\$ That is coming from Tamilnadu Electricity Board (Tamilnadu State Govermenment, India). I can stabilize each line before giving it to the load, instead I want to select the high voltage input and give it to the stabilizer, this way I can have only one stabilizer rather than three. (Stabilizer - Maintains constant voltage from a variable input voltage line) \$\endgroup\$ – Kannaiyan May 27 '13 at 20:35
  • \$\begingroup\$ Unfamiliar with how a "stabilizer" works or what it looks like, do you have a reference or an image? \$\endgroup\$ – jippie May 27 '13 at 20:47
  • \$\begingroup\$ Technical term would be Voltage Regulator. vguard.in/stabilizer/voltage-stabilizers/#!mainline_stabilizers . This is the one we use to stabilize the voltage. (billion $ business in India) \$\endgroup\$ – Kannaiyan May 28 '13 at 0:23

This seem to be good and cheap solution for the problem.

This seem to be a cheaper and reliable solution, without inductive circuits. Please suggest if I have missed something to consider. SMPS - Switch Mode Power Supply (5V).


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.