Correct grounding technique for Geode seismographs

Hi Anton,

We are not aware of any special grounding schemes for a large number of modules.

In dealing with EMI interference, our experience is that you need to individually ground each Geode - how you do it depends on the conditions you are in, but generally a Metal stake with a wire to the grounding post is the best you can do.  Sometimes adding a little water to the soil around the stake can help. 

We have seen a majority of this type of noise occur in highly resistive environments (dry sands or hard rock), and grounding can help.  If you are able to orient the lines perpendicular to the transmission lines that can also help.

In general, with high voltage lines, the emanating field (50 or 60Hz) is likely coupling into the spread cable. Each geophone pair in the spread cable is a big antenna, but the pairs are twisted so any signal induce in one will also be induced (in varying degrees) into its neighboring wire. This will lead to a common mode voltage at the input to each Geode channel. So it would depend upon the common mode rejection (CMR) of the Geode’s input amplifier to reject this common mode signal. But the CMR of an amplifier is finite, and if the common mode signal is large, the common mode range of a given channel can be exceeded leading to the 50 or 60 signal being amplified by the front end differentially resulting in a lot of noise. This can happen even with very good CMR.  Also any unevenness in the spread cable pair twists will result in a differential signal which will be amplified by the system.

We would also suggest trying the lower gain setting, in case the front end is being over driven. You may need a bigger source to overcome the loss of amplification, but if you can get data, even with substantial 50 60z noise, a post processing 50 or 60Hz filter may be able to clean up the record enough to be useable.

We haven’t tried this, but if everything else failed then this might be worth trying:

It is possible that the noise is  being induced in to the coils of the geophones. If that’s the case, then putting a metal can over the affected geophones might help. Note that the cans might have to be grounded with a metal stake driven into the earth.

Thank you for detailed answer.

I understood that common practice is separate grounding of each module. Also we will definitely try to shield each geophone with a separate ground, but this will require a large amount of preliminary work.

Regarding 50 Hz filtering.

We conducted a small experiment in the field to determine the frequencies of background electrical noise captured by geophones. Having set the minimum possible sampling period for Geod, I recorded 9 consecutive intervals (128 seconds each) of passive observations. This resulted in a total of 19.2 minutes of continuous background noise recording at 125 Hz sampling rate.

Then I calculated the signal spectrum from each geophone separately (19.2 minutes of recording), and then obtained the average spectrum for all sensors. For example, the figure in the appendix shows the average spectrum from 24 sensors (red curve) versus the spectrum of an individual geophone (blue curve).

In general, I am interested in noise in the 30-50 Hz range. There are no clear peaks at 50 Hz on the frequency response graph. Peaks at 20, 25, 30 Hz are present constantly at any time of the day. In general, the frequency response of passive observation signals increases smoothly with a maximum around 48 Hz.

I mean that filtering in some narrow frequency range in this case will not help much in my understanding.

Thanks again for the advice. We’ll experiment.

@antonydublin Thanks for sharing the plot! If the noise is broad-band as shown in the plot, narrow frequency filtering won't help much.