Acquire RVDT

[Ayous]

Hello all,

My company and I are trying to find a replacement solution of the SET cRIO RDK 9316 card, which acquire LVDT and RVDT sensors. Those cards will be no longer available soon so we have to find a good solution.

For LVDT we have no problem using converters, but for RVDT application it's a bit difficult, we can't find converters which converts linearly the 0-360° angle from our sensor to -x to x pts/V... as the RDK 9316 card did (0° -> -30000pts / 360° -> +30000pts for the 9316 RDK card). The converters we tried were not linear and reaches a maximum tension at 180° before returning to their initial value, so we had 0° -> -5V , 180° -> 5V , 360° -> -5V.

Does someone knows how the RVDT signal was demodulated by the 9316 RDK card, or does anyone has already found a replacement solution ?

Respectfully,

Le Meute Simon

[GerdW]

Hi Ayous,

---

@Ayous wrote:

The converters we tried were not linear and reaches a maximum tension at 180° before returning to their initial value, so we had 0° -> -5V , 180° -> 5V , 360° -> -5V.

---

This sounds like expected behaviour when I read this Wikipedia article…

---

@Ayous wrote:

Does someone knows how the RVDT signal was demodulated by the 9316 RDK card

---

I guess they look at both voltages V1 and V2 to determine the quadrant, then they unwrap that sawtooth signal (from Wikipedia) to a signal rising from 0° to 360°…

---

@Ayous wrote:

does anyone has already found a replacement solution ?

---

Use a simple AI card, sample the voltages with high sample rates and do all calculations on your own.

This is how I analyze the signals of our resolvers…

[Ayous]

Hi GerdW,

Firstly thank you for your fast and complete answer,

I thought to not use a demodulator and do the calculations with the FPGA but there is one thing I don't understand:

For a level x of tension, how can I know if I'm at the y1 position or the y2 position.

This is the statement I've done with my cRIO:

Effectively, if I only acquire Vs1 and Vs2, I really don't understand how can I demodulate the blue and orange signal to make it linear ?

Respectfully,

Le Meute Simon

[Jay14159265]

Why not arcsin of the grey data points?

[Henrik_Volkers]

one generator plus two AI channels sample the AC voltages, Frequency is known , measure amplitude and phase(!) of both signals do the math (demodulation)

you can try it with your soundcard (adding a input protection , 2 R plus 2 LEDs )

not an industrial solution, that would need a generator and AIs..

[Ayous]

Hi Jay,

The grey data points is the output of my demodulator which I don't understand how to copy the behaviour,

Moreover, if I arcsin those points do I have to divide all the values to set them under 1, then arcsin, and then re multiply by the scale factor ?

With this method I have this :

The shape looks good but I don't know if the method is correct,

Respectfully,

[Ayous]

Hi Henrik,

Thanks for your help, concerning the excitation, I have an exciting signal perfectly sent, concerning the acquisition, I think I measure the amplitude with an AI but the phase, how can I measure it with my cRIO ? And then, the question is the math, what should I use ? There is a lot of different version (arcsin/ arctan/ dividing, substracting) but I don't know which formula I need to follow,

Respectfully,

[Henrik_Volkers]

Do you have at least one type of your RVDT, where there is a datasheet?

Link to datasheet?

You need to measure the phase between the exitation and the RVDT output. So your DAQ needs to measure the exitation and the output at the same time (simultanious). 

In simple case, capture 20-50 periodes at a samplerate with least 10 (better 50 or more) points per periode.

Use extract tone information.vi to get amplitudes and phases. Amplitude ratio ist your slope and the phase difference (about 0° or about 180°) tell you which side of the V-slopes you are.  Around zero (ratio)  the phase difference will usually slowly change from 0° to 180° .. depending on the quality of the sensor.

[Ayous]

Henrik,

Thank you for helping me,

The characteristics of the RVDT is as it follows, it's made by our customer :

To summarize:

- I need to measure the excitation and the RVDT output at the same time (one AI each), i'll do it in the FPGA

- I convert it into signal labview data

- Thanks to Extract tone information.vi, I get amplitude and phase

- With the ratio I know the level of what I need to convert into degrees (°) and with the phase difference if i'm on the left or on the right of the V-slope, so I know if I need to take the conversion as it is or conversion = ( minimum of the V-slope - conversion)

Am I correct ?

Question: the RVDT output is VS1-VS2 or the ouput of the demodulator (cf. grey curve in my last message) ?

Respectfully

[Jay14159265]

---

@Ayous wrote:

Hi Jay,

The grey data points is the output of my demodulator which I don't understand how to copy the behaviour,

Moreover, if I arcsin those points do I have to divide all the values to set them under 1, then arcsin, and then re multiply by the scale factor ?

With this method I have this :

The shape looks good but I don't know if the method is correct,

Respectfully,

---

Well without knowing anything about the modules internal workings, if you can linearize the output over the range of interest, you can calibrate the output to give you the measurement you are looking for. If you have a linear output for the valid range of the form Y=mX+b where X is the input from the sensor (in this case linearized with the arcsin) and Y is the desired output. Use m and b to map X to Y. Rotate it to the maximum position and use a reference sensor to get your initial x1 and y1 point then rotate it to the other maximum valid position to get x2 y2 draw a line through those two points and you have your linear calibration line.