noise in usb-6221

[zqhector]

Hello,

I'm trying to take very accurate measurements using a DAQ board.

Right now, I have a USB-6009 (powered from the USB connection). When I connect it to take a measurement, it adds a lot of noise to the signal I want to measure. I see these behaviors while minitoring said signal with a scope. Then, I traced the noise back to the USB power from the PC. As it is right now, the random noise being added to my signal is  +/- 25 mV and it affects a great deal my end application.

Now, I'm looking at the manuals of another board, the USB-6221 (externally powered). And my question to you guys is, what noise figure (in mV) would I see with the USB-6221 on its power rails +5V and GNDs? My guess is since it is externally powered, it must have on board regulators: LDOs ? DC-to-DC converters ? filters ?

Thank you,

Hector

[SnowMule]

25mV is teeny-tiny with the 600x's.  Those are cheap inexpensive DAQs, not high-performance DAQs.

One trick is to take a few samples (100 samples at 1khz) and average them, then use that value.  Obviously this drops effective sampling rate significantly. 

Other option is to amplify the signal in hardware before you pump it into the DAQ, then divide the voltage back down in software.  I do this with current measurements since the voltage drop across a small resistor is tiny.  (I average readings with that too, and even after that there's still a lot of noise on the measurement.)

What's the range on the signal you're trying to measure?  The 600x isn't a *bad* DAQ, but you have to realize they're not built for high-performance applications.

Edit: Also, are you using differential or single-ended mode on the input?

[zqhector]

The range is from 3.0 V to 1.8 V.

The maximum time this voltage can be present to be measured is 5 ms. Therefore, I cannot take many samples.

I'm using single-ended inputs.

The main problem I have with the 6009 is that it injects noise: Around +/-25 mV max, randomly, on top of the signal I want to measure. This is caused by the noise from the USB power (PC side).

I think an externally powered board, like the USB-6221, might add much less noise to my signal, but I'm not sure about that. I was hoping you guys might know the magnitude of the ripple on the +5V connector of the USB-6221.

Thank you

[johnsold]

Computers and devices deriving power from them are always noisy.  No one designs a computer to be part of a high quality instrumentation system.

If you want the best performance, you will need to go to a measurement system that is not powered from the computer.  PXI may be a good option although it is significantly more expensive than the USB devices.  Consider how much time you will be spending trying to get an inadequate device working versus what it would cost for an appropriate device.

Asking your NI Field Sales rep for suggestions might be worthwhile.

Lynn