LVDT max resolution

[MateoAerny]

Hi everybody

For a project, I have to measure de deplacement of a XY table. The resolution of a mesure shoud not be higher than 0.1um (100nm). I'm projecting to use a DC-DC LVDT probe (Honeywell DLF-HQ) with an NI 9209 DAQ (24 bits resolution).

As the range of the LVDT is 50.4mm, I could theoricaly rech a resolution of 3nm as long as I use all the range of the captor. (50.4/2^24). I saw on the DAQ datasheet that the input noise of the card is 20 uV. I know that this noise will make my 3nm resolution impossible to reach but I would like to know how to calculate the effect of the noise on the final resolution.

Regards

Mateo 

[Henrik_Volkers · Accepted solution]

You want ~2ppm resolution!

How temperatur stable is your setup?

LVDT datasheet:

gain ~100ppm/K

zero ~300ppm/K

(not counting the material (steel) length properties;) )

9209

gain ~25ppm/K    ~50µV/K at 2V

Offset ~2.4µV/K

if you can make your measurements inbetween a 10mK temperature change ... maybe 😉

Noise

you need about 7.9µV resolution (4V/50.4mm*0.1µm)  if you mean 4 times you can expect half the noise, say 20 measurments (~1sec) will bring you in that range.

But if the 10µV of the noise is due to random walk of the internal reference it will not help you.

How big is the noise of the LVDT?? (what about cable (antenna) EMC noise)

I would go optical (encoder or interferometer ), frankly ... be happy if you reach stable 1 µm with residual random digits ...

[MateoAerny]

Hi Henrik

Thank you for the very specific answer. I will try to find anoser way to achieve the measurments (I saw optical encoders that can achive a 1um resolution  )

Regards

Mateo