Remove Phase Lag of Voltage Transformer Output

[shivakumar.r]

Hi Friends,

I working on Energy meter design which uses 1:1 transformer for voltage measurement and Current transformer for Current measurement.

While experimenting i found Voltage transformer's output signal lags by approximately 30 Degree with input signal.

Is there a way to remove this phase Lag ?
Some circuit or any SW algorithm to remove it.

Below is the circuit

 

[FvM]

I would rather expect leading phase due to transformer magnetizing inductance. The large phase error makes me wonder if the measurement also suffers from voltage non-linearity and waveform distortion?

Energy meter ICs usually provide a phase calibration range below 1 percent which seems an acceptable error for current and voltage transducers.

A correction is possible both in hardware (RC filter) or software (time delay for samples).

 

[shivakumar.r]

Thanks For Quick reply.

I cross checked the result.
Transformer output voltage is lagging.

Attached is the captured image for reference.
Red color signal is input(230V) and Yellow color is transformer output.

Can i get little more details on implementation on both HW and SW.

 

[FvM]

Transformer output is coming before input voltage. We call it leading phase.

Waveform is also slightly different from input (= distorted). That may be tolerable, but due to non-linear magnetizing inductance, you can also expect that the divider ratio changes with mains voltage variations. If you have a variac, you should check.

 

[KlausST]

Hi,

I wonder what transformer you use.
30° is way more than one can expect.
And the distortion may casue problems, too.

It it internally calculates U x I to get P, then the distortion will cause large errors.

Klaus

 

[shivakumar.r]

Thanks for the Reply..

Ok.
I understood about leading phase.

Mains voltage will be normally constant value.
I can manage with slight amplitude errors.

But due to this phase error, i am getting large error in Reactive power and Power factor.

I use custom made 1:1 transformer.
Transformer with higher accuracy costs more.
Which is not suitable for my project.

Simple voltage devider for voltage step down will not provide complete isolation from mains supply.

Is there any work around to reduce the phase error of transformer.
How do i compensate it.
Either by adding some extra simple HW or some work around in SW.

 

[schmitt trigger]

Indeed, 30 degrees is very significant!
Even if you had a purely resistive load, it would show as a PF = 0.86

In my experience that is caused by no having the proper burden resistor. Judging by the waveform amplitude, that is what it seems is going on. Please show your CT's datasheet and the recommended burden resistor.

But anyways, there will always be -even with unity power factor- a small discrepancy of a few degrees between the potential and current transformers. The image below is from a circuit I designed 20 years ago. It aligned the relative phases with an RC network on the potential transformer.

The 68 ohm burden resistor for the current transformer (T3) was obtained from the vendor's datasheet.

 

[crutschow]

The large resistor in series with the transformer primary is likely causing the problem.
You are seeing the phase-shift from the magnetizing inductance.
You need to use smaller resistor values, or use a 220V transformer with no resistors.

 

[FvM]

As the OP already conceded, the problem is caused by a transformer not suitable for accurate measurements in the given used circuit. Using a lower voltage resistance level would be a solution, but involves higher power dissipation in the HV series resistor.

A different possible solution could be to operate the transformer with virtually zero flux by a adding a transimpedance amplifier circuit with 1k feedback resistor.

 

[shivakumar.r]

Thanks For all the quick suggestions..

I will implement suggestions provided and update the results..

 

[schmitt trigger]

As I was typing my reply, the OP posted a response which includes he following statement, which I hadn't read:

"I use custom made 1:1 transformer."

This is exactly the problem. An improperly designed current transformer.

 

[crutschow]

As I was typing my reply, the OP posted a response which includes he following statement, which I hadn't read:

"I use custom made 1:1 transformer."

This is exactly the problem. An improperly designed current transformer.

But the circuit he shows is for voltage, not current.

 

[schmitt trigger]

You are absolutely correct Cruts!
I hadn't seen his original schematic.

He is showing a voltage transformer in which he intends to reduce the primary voltage solely with a resistor!!!

Your reply:

"The large resistor in series with the transformer primary is likely causing the problem.
You are seeing the phase-shift from the magnetizing inductance.
You need to use smaller resistor values, or use a 220V transformer with no resistors."

is essentially correct.

 

[crutschow]

..................
A different possible solution could be to operate the transformer with virtually zero flux by a adding a transimpedance amplifier circuit with 1k feedback resistor.

That should also work.
That way it is essentially being used as a current transformer.

 

[Warpspeed]

I suggest you use a commercial voltage transformer rather than try to make one yourself.
A suitable component might be a TV1013-1M made by Bingzi China.
Quoted as having less than 5 degrees phase error.

**broken link removed**

I just measured a test specimen I have here.
Both windings measure as 150 ohms and 11.4 Henries.

 

[shivakumar.r]

Thanks For the suggestions.

Please check the attachment for Transformer specifications.
Passing 1m Amps to transformer input and connecting 1K ohms at the output is correct ?

or can i use some other value ?

 

[Warpspeed]

Application notes suggest 2mA max working into no more than 250 ohms for non phase critical voltage only measurement applications.

Or better still, working into a virtual ground using an op amp, as suggested in the application notes for lowest possible phase shift, where phase shift is important.

 

[shivakumar.r]

Sorry..

I missed to attach my transformer specification.

Please check it.

 

[Warpspeed]

That should work reasonably well.
The lower the load impedance the longer the L/R time constant.

Best possible performance will be feeding it into the virtual ground of an an op amp, which has essentially zero input impedance.

 

[FvM]

For 1° phase error, the burden of the 3 H transformer must be as low as 17 ohm. Unfortunately the DC windings resistance also matters. Means you get phase error of 6.5° with 50 Hz/3 H/115 ohm even in a transimpedance circuit as suggested in post #9 and #19. According to specification, the transformer is simply not suited for precisison voltage measurements.

If you can get hold of a transformer with two secondary windings, or a tapped secondary, you can implemented a zero flux circuit with separate source and sense which also eliminates the winding resistance effect.