Feedback loop for constant voltage amplifier. Simulated Aircraft Power.

[Andytester]

Hi All, I have a project where I'd like to simulate Aircraft Instrument power at 115v 400Hz.
I've had success driving an amplifier from a Frequency Generator but if the load varies obviously the output voltage varies as well.
Is there a simple feedback loop circuit I could use to tap off the output and back into the input to control the final voltage at the test point?
A friend has suggested some sort of divider circuit with resistors but I cant find any ideas online as yet.

The load is only 20mA on standby and about 200mA when running.

Any assistance would be welcome, I've attached a simple sketch of the test layout.

Cheers

 

[Alec_t]

Does either the amp or the function generator have a designated gain/amplitude control input to accept a feedback signal?
Are the amp and FG commercial units or home brew?

 

[hevans1944]

To control the AC output amplitude you will have to control the amplitude of the signal generator input to the power amplifier. That generally means rectifying and filtering the AC output to create a DC feedback signal. You subtract this DC signal from a DC reference (control) signal to create an error signal. The error signal is then integrated and used to control one input of a multiplier circuit. The other input of the multiplier is your signal generator and the output of the multiplier drives your power amplifier.

An inexpensive four-quadrant multiplier with differential inputs and a single-ended ±10 V output is the Analog Devices AD633, which would be suitable for your application. There are other integrated circuit multipliers available, but it is easiest to work with a four-quadrant multiplier. You need at least a two-quadrant multiplier that accepts bi-polar input from the function generator and produces a bi-polar output.

 

[Andytester]

Alec, The FG is a basic Analogue one with no input for feedback, I guess I could look out for a more advanced unit. The Amp was a commercial 200w unit borrowed from a friend.
I was hoping to avoid spending £2000 on an AC power supply like the B&K 9801.

 

[Andytester]

To control the AC output amplitude you will have to control the amplitude of the signal generator input to the power amplifier. That generally means rectifying and filtering the AC output to create a DC feedback signal. You subtract this DC signal from a DC reference (control) signal to create an error signal. The error signal is then integrated and used to control one input of a multiplier circuit. The other input of the multiplier is your signal generator and the output of the multiplier drives your power amplifier.

An inexpensive four-quadrant multiplier with differential inputs and a single-ended ±10 V output is the Analog Devices AD633, which would be suitable for your application. There are other integrated circuit multipliers available, but it is easiest to work with a four-quadrant multiplier. You need at least a two-quadrant multiplier that accepts bi-polar input from the function generator and produces a bi-polar output.

Thanks hevans1944, I may look at the AD633 option if I can work out how to use it.

 

[hevans1944]

Thanks hevans1944, I may look at the AD633 option if I can work out how to use it.

For your application, controlling the output amplitude of a 400 Hz power amplifier, driven by a 400 Hz function generator, by means of a negative feedback control system, you would have to control either (1) the gain of the power amplifier or (2) the amplitude of the function generator signal. While you could possibly use a motor-driven potentiometer for either control function, it is much better and simpler to use a voltage-controlled, variable, electronic attenuator. That is the function provided by the AD633 four-quadrant multiplier. You apply an AC signal to its X1 and X2 differential inputs and a DC control signal to its Y1 and Y2 differential inputs. The W output will be the AC signal applied to the X inputs but with a variable amplitude proportional to the DC signal applied to the Y inputs.

You can use the AD633 at the front-end of the power amplifier to control the amplitude of the function generator output with a DC negative feedback signal derived from the power amplifier 115 V AC 400 Hz output.

It's pretty simple. You connect the function generator output to one pair of the multiplier differential input terminals. You connect the DC error amplifier and integrator output to the other pair of input terminals. As a result, the multiplier output will be the function generator output multiplied by a DC feedback voltage and a fixed scaling constant. For example, zero to ten volts DC can result in an output AC waveform that varies from zero amplitude to whatever the function generator output amplitude is. You therefore have created a "gain" adjustment that is controlled by a DC voltage. If the AC output voltage from the power amplifier rises above the set-point voltage of 115 V AC, the DC voltage is reduced to bring it back down to 115 V AC. If the the AC output voltage from the power amplifier decreases below the set-point voltage of 115 V AC, the DC voltage is increased to bring it back up to 115 V AC.

You could do this yourself with an AC voltmeter connected to the output, a potentiometer connected as a variable attenuator to the power amplifier input, your steady and quick hand in conjunction with your Mark I Eyeball adjusting the pot knob to keep the output voltage at 115 V AC. Nothing wrong with that approach except for speed and reliability issues. And it does tie up a human being who might want to do other things. So let's automate with a negative feedback electronic control system.

You connect the multiplier output to the power amplifier input terminals in place of the function generator output. All you need for negative feedback is an attenuated sample of the AC output voltage, rectified and filtered to DC, to compare against a set-point voltage that will be used to adjust and determine the AC output voltage. A rectifier, followed by a voltage divider and capacitor filter, all paralleled on the AC output separate from your instrumentation load, will get your high-level 400 Hz 115 V AC voltage down to a low-voltage DC that is proportional to your AC output voltage. This DC voltage you can use for negative feedback with an op-amp error-amplifier and integrator. This is called a PI (Proportional Integral) negative feedback control loop. The integrator part of it ensures that the error between set-point requested output and actual 115 V AC output will eventually become zero,,, no error at all.

A key element for your PI controller is the multiplier, which acts as a voltage-programmed variable attenuator for the function generator 400 Hz signal applied to the power amplifier. The other key element is an op-amp configured as an error amplifier with an integrating function. It must also have a means (a diode) to prevent negative outputs from being applied as inputs to the multiplier. Negative going error signal outputs will occur when the AC output is greater than the set-point voltage, thereby requiring a reduction in the AC excitation to the power amplifier.

The multiplier doesn't care if its DC input operand is positive or negative. It will attenuate the AC input operand for either polarity, simply inverting the AC signal waveform for negative DC input operands. However, it cannot produce a less-than-zero AC excitation, so it is necessary to prevent the DC operand to the multiplier from changing polarity because that would result in positive feedback and a maximum uncontrolled AC output. The essence of the problem is this: the multiplier must vary the amplitude of the function generator signal from zero to some maximum value in response to a feedback error signal that will be a positive polarity if more voltage is required (less attenuation) or a negative polarity if less voltage is required (more attenuation). The integrator function will eventually produce a positive control signal at a level appropriate to maintain the AC output at the set-point value, but a large negative error signal must be suppressed until that happens.

If you need some help designing and building the negative feedback control system, there are certainly lots of folks here who can help you.

Hop

 

[BobK]

Wait a minute. You have a 200W amplifier and it's voltage output drops when supplying 200mA? Something is wrong here. 117V at 200mA is only 23W. The amplifier should be able to handle this easily.

Edit: Okay, I see that there is a transformer in the output. I suspect the transformer is the problem.
What is the ratio? Perhaps too much current is required on the input to get 200mA out.

Bob

 

[hevans1944]

Maybe that 200 watt wall-banger amp was designed to drive the 4 to 16 ohms load impedance of a loudspeaker. If you hook up a 400 Hz step-up transformer to it, the amp may not like an open circuit on the transformer secondary. Try adding a permanent resistive load to dissipate a few watts, say five to twenty watts, all the time. If the transformer is big enough, the amp probably won't notice when you switch in your instrumentation load. You may not need output voltage regulation after all.

 

[Andytester]

Thanks guys. I've found a work around for now that uses a Variac after the fixed Transformer and we adjust that to balance out any small changes during the test. Its a bit cumbersome but it works.