How can I use DSP to analyze a noisy sensor signal?

[HappMatt]

First off let me just say that I am not a EE and have never taken a signals course before though I am sort of willingly having to over come this for a work project. I am working with a sensor that is giving me a noisy signal and I am trying to figure out if there is any usable data in my data. Of course noise reduction via all the usual methods have been done(grounding, twisting,..ect) other than DSP. The issue I am having is that I have something that is switching at 10hz and a signal that is likely at 10hz. I think that the signals phase is likly shifted and would like to try and look for that some how.

I am looking at various book trying to figure things out and was curious to try and make a type of graph where you plot time on the x, frequency on y and then phase using color. I have seen a plot like this before but with magnitude and don't know the exact name for it I thought is was a spectral frequency something. I am not sure if this will be the ticket or not but if there is someone who knows DSP better than me(which I assume is everyone) and can provide any guidance on this situation that woud be much appreciated.

Thanks

 

[berkeman]

First off let me just say that I am not a EE and have never taken a signals course before though I am sort of willingly having to over come this for a work project. I am working with a sensor that is giving me a noisy signal and I am trying to figure out if there is any usable data in my data. Of course noise reduction via all the usual methods have been done(grounding, twisting,..ect) other than DSP. The issue I am having is that I have something that is switching at 10hz and a signal that is likely at 10hz. I think that the signals phase is likly shifted and would like to try and look for that some how.

I am looking at various book trying to figure things out and was curious to try and make a type of graph where you plot time on the x, frequency on y and then phase using color. I have seen a plot like this before but with magnitude and don't know the exact name for it I thought is was a spectral frequency something. I am not sure if this will be the ticket or not but if there is someone who knows DSP better than me(which I assume is everyone) and can provide any guidance on this situation that woud be much appreciated.

Thanks

Can you say a bit more about the sensor, the cabling, and the environment? There could be some more things that can be done to improve your initial SNR. Can you show us a sketch of the setup?

 

[uart]

I am looking at various book trying to figure things out and was curious to try and make a type of graph where you plot time on the x, frequency on y and then phase using color. I have seen a plot like this before but with magnitude and don't know the exact name for it I thought is was a spectral frequency something. I am not sure if this will be the ticket or not but if there is someone who knows DSP better than me(which I assume is everyone) and can provide any guidance on this situation that woud be much appreciated.

Thanks

Yes a Discrete Fourier Transform (DFT) will give you the frequency and phase information. That sounds like a good way to go.

Can you tell us more about the signal. At such a low frequency of 10Hz there shouldn't be too much noise there. You might get a bunch of noise at the mains frequency (60 Hz) and then some broadband "fuzz" as well. But if you can band limit to around 10Hz I'd be surprised if there was a lot of electrical noise there. Unless this "noise" is unwanted signal, say or example acoustic noise from the same transducer which your signal is coming from.

 

[HappMatt]

Sorry for the delay in responding, thing are a bit busy. A quick description of what I can say about the system is its a magnetics based sensing system consisting of a magnetic sensor(cant really say more), Labview with a cDAQ and a switching device that controls the rate at which samples are seen by the sensor(sorry, I think I need to be somewhat vague). So for the setup I have everything taped down and stuff grounded and removed all other electrical devices that are in the vicinity or on the same outlet, some of the cables are shielded and those that are not are twisted( its a experimental set up thus there are unshielded wires that sort of have to be there). There is a signal generator which is switching at the 10hz frequency and a battery powered motor. I have been working on moving the switch farther from the sensor and the motor farther away which I am hoping helps. There was also another realization that was had the other day which is that even though the samples arrive at the sensor at 10hz, the samples velocity is actually faster and thus what I should see is a wave appearing at 10hz intervals but that have a frequency of ??( I have to find the number again but I think it was around 150hz from a calc). So I am getting things set up again and will check to see how my SNR is, if I can find the signal. I have also picked up a couple books on DSP and and gradually going through them at evening. I have access to MATLAB and LABVIEW and will try a DFT of the signal once I get new data and let you know what happens. If there are any suggestions, they are gladly accepted.
Thanks