Measure Tiny Signals Buried in Noise on Your Oscilloscope (Phase Sensitive Detection)

by andrei101 in Circuits > Electronics

3523 Views, 13 Favorites, 0 Comments

Measure Tiny Signals Buried in Noise on Your Oscilloscope (Phase Sensitive Detection)

picture1001.jpg
Full vid

Imagine you want to measure a tiny signal buried in noise that is much stronger. Check out the video for a quick run through on how to do it, or carry on reading for the details.

Example

pic1.jpg
Schematic_vid-pd-amp_Sheet-1_20181028101230.png

Imagine you want to measure the light reflected from a laser spot using only a photo diode with no optics and a crude amplifier.

You can see the signal that we get is dominated by the room lights as well as the 50 Hz noise picked up by the amp.

Simply averaging your signal will not work here as the background changes (say you moved your hand) is far more significant the the effect of blocking out the laser to measure the difference.

This is a terrible setup because you are trying to measure a signal at DC, and this is a very noisy area of the spectrum. But as you go further into the AC the noise generally decreases because the main source of noise is called pink noise: www.wikipedia.org/wiki/Pink_noise

So the solution is to move our signal into the AC, away from the noise sources.

Solution

pic2.jpg
SDS00003.BMP

You can move the signal into the AC by pulsing the laser, and the way i've done that here is by powering it from a digital pin on the arduino. The arduino is running a blink sketch which makes a 5khz square wave to power the laser directly.

you can then hook up another probe onto this pin to tell the oscilloscope the exact frequency of the laser.

Now that the signal is in the AC you can AC couple channel 1 to get rid of the dc offset and maximise the dynamic range of the ADC.

Then you want to set the trigger for channel 2 as this will be the exact same frequency as the light emitted from the laser.

Now we can see that there is a tiny square wave in the noise. This is the light from the laser!

And because we are triggering at the same frequency we can average out the signal: anything that is not the same frequency as our signal, or random noise, will average out to 0.

Our signal which is always in phase with the reference channel will average out to a constant waveform.

Results

SDS00004.BMP
pic3.jpg
SDS00005.BMP

You can see that we have dug out our signal from all that noise! this is essential making a band pass filter which gets narrower as you include more averages.

The signal is around 50 mV and it was buried in 1 V (peak to peak) of noise! amazing that we can still measure it!

The result can be justified by blocking out the laser which forces the signal to disappear.

This technique is called phase sensitive detection and has many uses, for one it is pretty much the backbone for all RF communication in the world!.

There are instrument called lock in amplifiers which can extract nV signals buried in V's of noise using this method. For a more comprehensive explanation and for ways to build circuits using this have a look at this analog devices article:

https://www.analog.com/en/analog-dialogue/articles...

I hope you have enjoyed this quick hack, if you have any questions i'd be happy to answer them in the comments.

If you found this useful you can give me a vote :)