DAQ voltage inputs affecting each other and the device they're recording from

[rachgilby]

Hi everyone!

I am using ELVIS II+ and LabVIEW 2012 to record voltage from a device with a resistive sensor (See Voltage measurement with ELVIS II.jpg for VI). The device (Device+ELVIS II Schematic.jpg) measures the voltage across the resistive sensor and is controlled by an msp430. If the msp430 finds that the change in the sensor's voltage is greater than the threshold, an LED is turned on. There is a motor function on the device and the LED is normally not on when the motor is on. I have found that recording with the daq affects the behavior of the device. The device and daq grounds are connected. When the daq is NOT recording, the sensor raw and analog in voltages are constant (and the LED off) while the motor is on (Figures.doc; A). When the daq IS recording, the motor voltage seems to distort analog in/sensor raw voltage causing the LED to turn on (Figures.doc; B). 

Tests performed that did not fix problem (LED on while motor on): 

1. Recording with new ELVIS II+ DAQ

2. Physically separating sensor raw and motor inputs on the DAQ

Tests performed that did fix problem (LED off while motor on):

1. Grounding sensor raw and/or motor

2. Separating sensor raw and motor channels in DAQ assistant (Figures.doc; C: shows motor and sensor raw separated by A_VCC. Figures.doc; 😧 where motor and sensor raw are next to each other). I am not sure why the order of the channels in DAQ assistant would affect the behavior of the sensor signal.

Any advice on the root cause of this problem and ways to troubleshoot/potential long-term solutions would be much appreciated!

Thanks!

[Wan-L]

It sounds like you already found a fix for your problem.

Have a look at this article on grounding considerations, hopefully that clarifies some confusion

http://www.ni.com/white-paper/5362/en/

[rachgilby]

Thank you for your reply Wan. I realize that I was a little unclear when I said that I grounded motor/sensor raw. I'm sampling them differentially and the negative input was already grounded. When I said that I grounded the signal, I grounded the postive input so essentially I stopped recording from motor or sensor raw. While sensor raw is not completely necessary, I would like to continue recording it if there is another solution available.

Thanks!

[rachgilby]

http://forums.ni.com/t5/Multifunction-DAQ/NI-USB-6218-DAQ-Assistant-Scan-Order-Affects-AI-Signal/td-...

I think that ghosting may be the problem. It still seems strange to me that the order of the channels in DAQ assistant affects this and that the order of the physical channels on the DAQ doesn't seem to have an effect.

[Wan-L · Accepted solution]

If your device does not have simulteanous acquistion, then each channel is multiplexed into the ADC. There is input capacitance inside your device, which gives it some "memory". If the voltage stored in the capacitor is not discharged from one scan before the next scan starts, then the second scan will "remember" results from the first scan, giving you really weird problems. The higher the source impedence on the current channel th e greater the likelihood that the signal from the previous will be "remembered". Thats why the order matters, the highest impedence source goes first. 

[rachgilby]

Thank you for your explanation Wan! The DAQ is multiplexing so that makes sense.