AGC without phase shift

[sanwal209]

Hi,

I am looking for a circuit or IC which takes 0.5Vp-p to 1Vp-p sinewave and it outputs sinewave 1Vp-p without phase shift.

Background is we need to downgrade mains AC 110V to 220V to constant voltage sinewave without phase shift.

Best Regards
Sanwal Saleem

[Audioguru]

An ordinary AGC circuit using a Jfet produces no phase shift.

[sanwal209]

Can you please refer me to some schematics?

[DaJMasta]

Pretty much any analog AGC will do the job, right?  And if your signal is only 50Hz or 60Hz, your tolerance for phase shift is MASSIVE before anything is noticeable, so virtually any AGC will work.

[sanwal209]

I have tried to simulate couple of AGC in Tina TI but it does introduce phase shift

[Zero999]

Any active circuitry will produce a phase shift because there will be some delay between the input and output voltages changing, as nothing can happen instantly.

[Marco]

Just do it digitally ... doing this with a non sampling analogue circuit is going to require a complex circuit. Doing peak detection in discrete circuitry to tweak an AGC can be done, but at that point you're so much more components than a microcontroller.

Just use a microcontroller.

[Zero999]

Just do it digitally ... doing this with a non sampling analogue circuit is going to require a complex circuit. Doing peak detection in discrete circuitry to tweak an AGC can be done, but at that point you're so much more components than a microcontroller.

Just use a microcontroller.

But that would most likely result in even greater phase shift, than an AGC.

[Marco]

The phase shift from ADC/processing/DAC is almost irrelevant on a 50/60 Hz signal. Also this is almost certainly just part of a chain of processing, he needs to do all of it digitally.

This stuff doesn't make sense any more to do in analogue other than a mental exercise.

[SiliconWizard]

Knowing nothing more about the OP's requirements, it's hard to suggest anything really relevant.

We don't even know how they attenuate the mains signal? Is it direct coupling, with just a resistive divider, and no isolation? Is there AC coupling? Is there a transformer? All of this will already introduce some amount of phase shift. We don't know if that matters or not. We don't know what amount of phase shift is acceptable. As already mentioned, it's not that hard to get negligible phase shift on a such a low frequency signal as long as you use a reasonable topology (for instance avoiding the use of filters with cutoff frequencies that are too close to the original signal's frequency). We also don't know why they'd need a fixed 1Vpp output signal, or if they could actually make use of an output signal with an amplitude the dynamic range of which seems pretty small (6dB!)

So maybe they should post a schematic and figures about what they expect first, and how they intend on using the output signal.

[mikerj]

I have tried to simulate couple of AGC in Tina TI but it does introduce phase shift

Show your circuit.  Do you really mean you are seeing significant phase shift between input and output signals, or are you actually talking about the response time of the AGC circuit?  A simple JFET AGC circuit should introduce negligible phase change at such low frequency.

[sanwal209]

Hi Guys,

Sorry for late response, Let me explain the requirements.

Main scope of project is to convert single phase to three phase output. Output should be sync with the input phase. Yes, I know we can simply make inverter and do the job but our output current requirment is 1mA@220V. With this power rating inverter will be an over kill. Output is used to excite the coils so that other sensor board will be able sense.

Solution i thought is to downgrade 220Vmains to 1V(or less) fed it into two phase shift register and i result we will have low voltage 3 phase output which then fed into power amplifier and output of power amplifier will drive transformer to create isolated 3phase output.

AGC is required to keep constant voltage output after downgrading 220V mains so that any fluctuation will not effect the output.

I have attached some circuits which i simulated and got the phase shift.

[Jeroen3]

You can do this digitally. Just use a phase locked loop with a DAC waveform generator.
https://dsp.stackexchange.com/questions/52981/synchronize-a-software-waveform-generator

[sanwal209]

Yes, Everything can be done on digital level but we want to do it in analog as we dont have resources for firmware development.

[Marco]

Opamp peak detector -> low pass filter -> drive the output of the low pass filter every so slowly to 1V by changing the gate voltage of a JFET variable attenuator with an opamp circuit, negative feedback and very very low bandwidth.

This will have the response time on the order of seconds and there's no real way to avoid it without going digital, no phase shift though.

[Zero999]

I have attached some circuits which i simulated and got the phase shift.

How much phase shift does the simulator show?

Please post the results.

[Audioguru]

The Wien Bridge oscillator has only one output so it does not have a second signal to compare with its output phase.

[Zero999]

The Wien Bridge oscillator has only one output so it does not have a second signal to compare with its output phase.

I presume the idea is not to build the oscillator part, just the AGC. R2, R2, C1 & C2 will be omitted and the signal applied to the non-inverting input of the op-amp.

[Zero999]

Awhile ago I designed a microphone AGC for a project which got cancelled. Once the gain had stabilised, there was minimal distortion and phase shift, other than the usual 180^o expected from an inverting amplifier.

Perhaps the original poster has selected the wrong value for an AC coupling capacitor.