Hi, I am working on a project to take an input AC voltage , Vi, and keep it's output, Vo, held at a fixed 5.00V. I've studied a few ways to achieve this using a few AGC circuits. I am unable to get them to work properly and I believe it is due to my misunderstanding of the control loop. I'm using the AD633 Multiplier IC as a Varaible Gain Amplifier. I'm using the AD736 RMS-DC converter as my "detector" stage. And I'm using a difference amplifier as my error feedback. The transfer function of the AD633 as it is configured in my circuit is: (Vi*Verr)/10. For the sake of simplicity, I omitted the division from the transfer function, leaving the AD633 to just: Vi*Verr. I need my dynamic range to be quite wide: 5V - 7.5V.


My problem is that if I make my Vref = 6V, so that my error becomes 1 to satisfy a non zero output, the output isn't being adjusted to satisfy my dynamic range requirements as I change inputs. I plan to put peak limiting circuitry in on the output of this to correct the huge spikes occurring as AGC circuit is settling. How do I design this to the range that I need? Do you have any suggestions for how I can treat this more like a negative feedback control loop and solve mathematically for steady state error and such?

 

Post your .asc file.

 

Post your .asc file.

I've uploaded two different schematics I've been working on.

 

What's the reason for the RMS converter (which I don't have the model for)?
Is regulating to the waveform average value not accurate enough?

 

I've attached the model subcircuit here for the RMS-DC converter. I went the route of the RMS-DC converter because it is what I have in stock. I want to get accurate measurement of the input waveform to feed my error amplifier for comparison to a reference value. I've read I could use a peak detector circuit, averaging circuit, etc for the detector stage of the circuit, but this was what I had on hand.

 

Is this the circuit you are trying to use from the AD633 data sheet?
I suggest you try to simulate that exactly and then do any modifications you may need after you get it working.

 

Is this the circuit you are trying to use from the AD633 data sheet?
I suggest you try to simulate that exactly and then do any modifications you may need after you get it working.

View attachment 133978

Yes! That is one of them that I've tried to model, however I'm having trouble understanding how that circuit works. My understanding is that the AD633 will realize the following transfer function (Y1 * X1)/10. Then the output W goes to an OP AMP (A1) with an adjustable gain up around 10. This output then goes to the RMS-DC IC via a 1/11 voltage divider. That output then is fed to an Integrator (?) with a reference level adjust on it. What is the dynamic range of that circuit? I couldn't find it in the datasheet. Is it a PI control loop? I've tried modeling the above circuit and it produces a similar outcome to my simplified circuit.

 

Actually it's a negative feedback I control loop with no P.
The integrator (A2) will give zero error at the set point with no gain error.

The AD73 generates a DC output proportional the the AC signal from A1.
This signal is compared with the reference voltage from pot R5.
Any difference between the two causes the output of integrator A2 to change, which controls the effective gain of the AD633 multiplier to correct the output voltage level from A1 to the desired value.

The dynamic range of the circuit should be quite high since the gain of the AD633 can go from essentially zero to 1 for a maximum 10Vpk input voltage.

What exactly is the change in signal amplitude that you having a problem with?
How fast does this change occur?