Lock-in amplifier driving me crazy

[Topher925]

I have a project where I'm measuring the phase of a signal compared to a signal reference using an SR830 lock-in amplifier. My problem is that the amplitude of the input signal periodically changes over time and this is causing a false measurement in the phase of the input signal. I can't figure out what would cause it as the phase measurement should be constant and independent of amplitude, should it not? Keep in mind that this signal has a very good SNR.

The reference signal is a constant amplitude and frequency 50/50 square wave. It doesn't matter what filtering settings I have, if the amplitude of the input signal decreases by 50% the phase measurement will change by roughly 2 degrees or so. What is going on?

 

[M Quack]

If the amplitude is unstable, why are you so sure that the phase of your input signal is stable within 2 deg?

 

[NascentOxygen]

The reference signal is a constant amplitude and frequency 50/50 square wave. It doesn't matter what filtering settings I have, if the amplitude of the input signal decreases by 50% the phase measurement will change by roughly 2 degrees or so. What is going on?

Perhaps the method you are using to alter the amplitude (e.g., potentiometer) is also causing a change in phase? Maybe you are inadvertently changing the duty cycle, though one would expect the detector to be immune to this.

You will just have to do more testing to determine the basis for the apparent phase change.

 

[skeptic2]

What type of oscillator is your reference? I had a similar problem with purchased, ovenized crystal oscillators at 3 MHz. Our circuit compared the frequency of two such oscillators to detect long term drift. Like you, when the signal fluctuated in amplitude, we detected a shift in frequency. Since our oscillators were purchased, we simply sent them back to the manufacturer as defective to be replaced.

 

[Topher925]

I know that the phase is stable because I can measure it using an oscilloscope. However, I think I may have found the source of the problem, a second weaker signal (same frequency) mixed in with signal I'm trying to measure. I'm still investigating to see if this is the actual cause but given the magnitude of the phase shift I don't believe it is.

I'm actually just using a micro-controller as an oscillator reference. The signals I'm working with are rather low frequency (50-200Hz) and the uC provides enough accuracy for me. I've checked the uC against the internal function generator of the amplifier though and its definitely not the problem.