NI9220 Compact DAQ and LVDT DC Accuracy control

[ghattas.ak]

Hello

I am working with the DC-LVDT Omega 620-25

http://www.omega.com/pptst/LD620.html

Manual: http://www.omega.com/Manuals/manualpdf/M4345.pdf

Specs: http://www.omega.com/pressure/pdf/LD620.pdf

I am using the NI 9220 module on a compact DAQ for data acquisition. the module supports differential inputs and single ended mode.

NI 9220: http://www.ni.com/pdf/manuals/373920a_02.pdf

I am experience some trouble obtaining 3 digits of accuracy from the converted signal, the excitation voltage is 12.75 Volts (external adapter) and the reading is -5.9

Furthermore the second case was by using a single ended connection (as per the modules datasheet) single ended is achieved by connecting the signal - of the LVDT module to the COM input, I also placed a 100nF capacitor for noise filtering, I was able to get 2 digits of accuracy 1.12x.

Input is read with a continuous acquisition mode of 100 samples/second with a 500Hz frequency and a voltage range of -10 and 10 as device parameter options ( only differential mode is selectable)

In addition and since the LVDT is linear I should directly convert the voltage reading to distance but the output voltage is of different range the datasheet states that for an lvdt of -d/2 and d/2 it should outputs a -5 and 5 V readings.

I double checked on the sensitivity formula mv/V/mm Output Ripple: 0.02% FSO LD620-25 ±25 (±1.0)inch and with a range of -6 -> 6 the full LVDT range is (12V) therefore is its of 0.0024 mv/V * 12.75(excitation) = 0.0306mm

Example used: http://tinyurl.com/hrcshow

but i didn't use that as for calibration I am considering the initial read voltage to be the 0 reference and then subtracting it from each reading!

how can I eliminate additional noise (AC noise) that is affecting my measurement and what should I take into consideration for designing a filter and choosing it's type

when disconnecting the LVDT excitation voltage the modules outputs uV values that varies. considering the copper wire acts a receiver and generating additional noise I place the 100nF capacitor between the signal + and signal - terminals.

the datasheet of the 9220 suggest using a 1Mohm between AI- and COM but that didn't improve the reading and the LVDT suggested a 10Kohm resistor between SIG+ and SIG- but that did not improve it as well

I tried the ring and twisting the cables but that did not work either the only working part with 2 digits accuracy is using a single ended connection where AI+ to sig+ and AI- to COM to sig-