problems with of analog signal with daq 6009

[ManuSpadim]

I'm having a problem with my NIDAQ 6009. as you can see in the video below, I'm having a big variation of the analogue signal reading. I tried to change the phiscal channel but it didn't work. The sensor I'm trying to read is a laser that measures distance. The laser gives from 1 to 10v analogue output and measures from 200mm to 2000mm, so, each 5mV means 1mm, and I checked it with a multimeter and the laser seems to be giving this precision. I wish I could have at least 1 mm of precision, but the variation is being bigger than 5mm. I tried to read  an average of an amount of received signals, but it didn't work, the problem is not some eventual wrong reading, but lots of them. I'm using the 1 sample on demand mode.

so, hope you guys help me.

thank you all! 

http://www.youtube.com/watch?v=YBuIq_3Tfi8&context=C30f82aeADOEgsToPDskL8SKJlGaKcY0TQ1Z3KJuUL

[RavensFan]

It looks to me like you are picking up electrical noise in your analog signal.  The device is actually sitting pretty close to your laptop power cable, and laptop power supplies can be noisy.  So it is quite possible it is picking up some noise from that.

Instead of 1 sample on demand, acquire multiple samples at a higher acuisition rate and send it to a waveform chart.  See what the real signal looks like.  Send the waveform through an FFT function and see if you are picking up any frequencies.  I would suspect you are seeing 60 Hz noise  (or perhaps 50 Hz since you sound European).

If you can determine the source of the electrical noise, you can take steps to try to shield or eliminate it.  If you don't need a very rapid measurement update (such as you only need 1 reading per second), take multiple samples and average the readings together.

[BZajec]

Are you using single ended or differential analog input?     I think you should be using differential analog input.

best regards,

 Bojan

[ManuSpadim]

Hello Ravens.

I don't beleave it's an electrical  noise.

I tried to make it with the cable far (about 1,5 meters) from electrical wires, but it didn't work. And I cannot put it more distant than this.

Do you think it can be the device that not support? as I calculated my device shoul have a precision of 0,0006 volts, since it has an analogue input of 14bits, and can read 10 volts (10 / 2^14=0,0006). Am I right?

thank you! 

[ManuSpadim]

I'm using diferential analogue input

[SnowMule]

For sure use differential.

Shield the DAQ, cables, etc. 

Like Ravens Fan said, take a hundred or so samples and average them before displaying them.  I usually read 100 samples at 1kHz, average those 100 samples, then use that average as the value.  Takes your effective sample rate down to 10Hz, but it's a lot easier to see what's going on when you don't have all that noise there.

[RavensFan]

---

@ManuSpadim wrote:

Hello Ravens.

I don't believe it's an electrical  noise.

I tried to make it with the cable far (about 1.5 meters) from electrical wires, but it didn't work. And I cannot put it more distant than this.

Do you think it can be the device that not support? as I calculated my device shoul have a precision of 0.0006 volts, since it has an analogue input of 14bits, and can read 10 volts (10 / 2^14=00006). Am I right?

thank you! 

---

Don't be so quick to dismiss electrical noise.  It is pretty common to have electrical noise problems even if you don't believe you should.  Have you tried using a waveform chart to look at the data?  That can give you a good idea of the nature of the problem.

What does the number on screen of your video represent?  Is that 11-12 number representing millimeters?  You may be looking at bit fluctation because you are right down at the resolution of your DAQ device, even as good as it is at 14 bits.  You have a precision of 3 decimal places on that display, but that may be way below the resolution of your device.

Let's work the numbers.  You say you 5mV = 1 mm.  Your screen (if that 11-12 number is millimeters) is only fluctating about a few tenths above and below 12.  A few tenths of a millimeter would be about 1 mV.  First, if that 1 millivolt is actually electrical noise, you are actually doing very good.  And with the math you did above for resolution, you came up with each bit being 0.0006 V, or 0.6 mV, that is very close to the 1 mV value or a few tenths of a millimeter.

If your device is sized to read a range up to 2000 mm, it is a bit much to ask to expect it to read to a 1 mm resolution.  That is only 0.05%.  And to have have a visual display that goes into another 3 significant digits beyond that is just ridiculous.

My final conclusion, you are using the wrong device with too large of a range, or looking at too tiny of a resolution on that device than what it is intended for.

[ManuSpadim]

I'm working at this problem, and I'll share with you my results as fast as I understand it.

thankyou all.