Cross-talk related issues

[Ekaterina Skidchenko]

We work with Gen 1.0 sensors. Figure stick_together (attached) is an empty room recording of analogue output, where X axis – time in seconds, Y axis – signal in nT. There are zones on the plot (marked by circles) where outputs from all sensors or some of them stick together for some time. Could the cross-talk be a source of this problem? And how to overcome it? Sensors were placed quite far from each other (around 3 centimeters between housings).

Moreover, the output of one sensor reaches from time to time some kind of a border near -2 nT (marked with arrows in Figure border, attached). Could there be some other reasons aside sensor saturation caused by sensor moving in high-level ambient magnetic field, i.e. exceeding the dynamic range of magnetometer? We work without active shielding, but our sensors and subjects are fixed. 

Finally, as you can see on both figures signal changes not above and below zero, but around approximately -0.75 nT (Figure border). Can it be somehow connected with calibration values? How can we fix it? Thank you in advance. 

[James Osborne]

At a distance of 3 cm between housing we would not expect cross-talk to cause an issue like this. It is difficult to determine what is happening from just the plot. If recording analog data make sure to check that the cables from the analog outputs are making good connections. Another possibility is that the cell temperature is not stabilizing properly. Can you confirm if the cell temp LED (LED2) stays green during the measurement time?

The analog data from the sensors will saturate near +- 2 nT in the default 1x gain mode with the conversion factor of 2.7 V/nT. You can switch to 0.33x mode which has a conversion factor of 0.9 V/nT (there is a button in the QZFM UI or you can send command 97). This will stop the data saturation at +- 2 nT and allow 3x the measurement range to be recorded but the dynamic range does not change. 

There is always some offset from zero but -0.75 nT is more than expected. Running field zero again should bring the center point closer to zero. If there are large drifts in the ambient magnetic field it will be necessary to field zero more often. Does this help solve that issue?  

[Ekaterina Skidchenko]

What is the distance where cross-talk is expected to cause an issue like this? LED2 of some sensors was flashing from time to time during measurements. How can we fix that?

Gain change helped, thank you, but running field zero more often did not helped. I wonder if we do the field zeroing in a wrong way. In your tutorial video you say that we should press 'field zero off' when the compensation field values stabilized. Could you please clarify what you mean under 'stabilized'?

понедельник, 10 августа 2020 г., 20:24:33 UTC+3 пользователь James Osborne написал:

[James Osborne]

To clarify we would not expect cross-talk to cause an issue like this at all. LED2 is the cell temperature stabilization indicator and will change from green to yellow during field zeroing, calibration, when the sensor is moved, or when the ambient field changes by a large amount. If LED2 is continuously oscillating between green and yellow and does not stabilize to green then it will likely be necessary to modify the cell temperature PID gains with  firmware update. More details on when the LED2 changes from green to yellow and how often it happens would be helpful to determine if a firmware update is necessary. 

The stability of the field zero value will depend on the ambient magnetic field so it is difficult to place an exact value on what would be stable but generally the B0, By, Bz values will not change by more than 1nT after a few seconds. The more important indication of local zero field is if the calibration factor is good (calibration factors should generally be less than about 1.5 and lower values are better). Field zeroing will generally take between 5-10 seconds.