Is this normal for a USB-6008 device?

[wingswing]

How noisy should a signal throught USB-2008 be? I've attached a picture of a signal (differential voltage measurements). The blue represents the signal after some filtering through LabVIEW, and the red is the time-averaged signal. My question is whether this amount of noise is normal for this device, or if there is something I could do to improve the signal (and if there is, what can I do?).

Thank you! 

[AnalogKid2DigitalMan]

Hello wingswing:

Gotta love those arbitrons 🙂

The 6008 is the bottom of the barrel DAQ card. Was that screen shot showing the full range of the noisy signal? Looks to be ~0.5mV.

What input range are you using? The manual specs noise as 0.5mVrms on the +/-1V input range, it goes up to 5mVrms on the +/-20V range.

Also note in DIFF mode, you have only 12 bits of resolution. At best with the lowest input range of +/-1V, the resolution is 2V/4096 = 0.4mV.

Based on the above, the noise does look reasonable for the 6008.

What is the nature of the input signal? It looks pretty small in amplitude, are you using proper shielding and wiring practices?

Can you connect a pure DC source like a battery and repeat your test.

-AK2DM

Message Edited by AnalogKid2DigitalMan on 06-12-2009 08:07 AM

[wingswing]

Hello AnalogKid, thanks for your helpful replies. (Actually, now I realize those probably aren't arbitrary units. Oh well.)

My input range should be very small: maybe from 40mV to 70mV. So about 30mV. The screenshot is a close zoom-in on the signal With 12 bits of resolution, I believe that should give me approximately .01mV resolution, is this correct? Because this is in fact the resolution that I need. 

The measurements I will be making is at steady-state, so ideally the signal should be flat, to within less than .01mV.

The test I did in the screenshot was done with a decent power supply. I just tried it quickly with a 1.5V battery, and got a value of around 1.38V, but the signal fluctuates between 1379.1mV and 1381.2mV. I just can't seem to get a steady signal.

What are some proper wiring and sheilding techniques?

In any event, I don't see this device working for my application. Any thoughts?

[AnalogKid2DigitalMan]

The resolution is based on the setting of the input range for the DAQ, not the amplitude of your signal. See my example above for the +/-1V range.

Another example: If you are using the +/-10V input range, that is a span of 20V. 12 bits of resolution yields 2^12 (=4096) discrete levels for resolving a voltage. 20V/4096 = 0.00488V of resolution.

Unfortunately, the 6008 is not up to the task.

Somewhere on NI's site is a paper on proper wiring and shielding practices. Type those terms in the search box at the top, you should come across some good resources.

-AK2DM