How does short input noise level pertain to the NI 6225?

[etechstan]

I have incomplete documentation on my company's product and am trying to find what short-input noise level is and how it pertains to the NI 6225. The previous manufacturer referred to this stating that it will be an improvement because the previous daq was 12 channels and the new one (6225) is 16. Any assistance will be greatly appreciated.

[CWolf]

The short input noise level is the amount of noise you will see if you connect the input to ground, so that it should be reading 0V.  There are noise and accuracy specifications for the 6225 in the Specifications Document.  These are not specifically when the channel has been shorted to ground, but are valid for different ranges of the card. 

I do not understand how your question about the noise levels is related to the number of channels on the device.  The 6225 has 16 single-ended analog input channels (or 8 differential channels), and all of them have the same specs for noise and accuracy. 

I hope this helps you, but if you have more specific questions, please let me know.

-Christina

[etechstan]

Christina, as I stated in the post, I have incomplete documentation from the contract builder and because of legal reasons we cannot contact them so I am basically reverse engineering the product. The requirement states;"The unit test for SYSREQ56975 (Short Input Noise Level) has not been run, but through analysis, we anticipate that the results should be as good or better than the legacy system due to the 16 bit vs the 12 bit DAQ card".  I have to explain these thing to the FDA but can't find any proper explanations.

[etechstan]

Is there a method to measure noise at the DAQ?

[Seth_B.]

It sounds like they didn't actually test the noise on the system, but rather assumed that because one system had a resolution of 16-bits and the other of 12-bits that the 16-bit system would be more accurate.  This is generally the case, but not guaranteed.  To test this "Short Input Noise Level", my assumption is that you should short the channel to ground and read on that channel, to determine your noise level.  For instance, I took a 12-bit C Series module (that was about 4 years out of calibration), shorted Analog Input channel 1 to the ground pin and read data on that line.  I got these results, which indicate that, when grounded, I have about ±15 mV of noise on the line.  Without more documentation, I can't be certain, but I'm pretty sure that this is what they mean be "Short Input Noise Level".

Regards,

Message Edited by Seth B. on 10-16-2009 09:47 AM

[etechstan]

Is there any location on the 6225 to connect to an Oscope?

[Seth_B.]

Do you have a PCI or USB version?  And are we concerned about input or output noise level?

If it is input noise level (the amount of noise on a reading taken on a signal input to the DAQ card) then an ocilloscope will not be of assistance.  You'll need a precision function generator or a DC power supply.  If you are just concerned about the noise on a shorted channel, you won't even need that.

If you are concerned about noise on the analog output lines, then you should connect the oscilloscope to the AO and AO GND channels.  For specific pins, I'll need to know if you are using a USB or PCI device.  If it's PCI, I'll need to know what terminal block you are using.

Regards,

[etechstan]

Seth, Its the PCI version with a filter buffer connected at port 1. You can get an overview of the product at spinematrix.com. I do have an Agilent function generator.

[Seth_B.]

Hmmm.  I'm not sure what the filter buffer module you are using.  Is it an NI product or custom-made? If you want to verify the accuracy of the 6225, you'll want to connect your function generator to AI channel 0+ and either 0-, AI SENSE, or AI GND depending if you are measuring in differential, NRSE, or RSE mode.  Also, you'll need to take the accuracy of your function generator in to account.  For instance, if you are making a measurment in the ±1V range, then at full scale (reading a 1V signal) the accuracy should be about ±0.36 mV.  However, if your function generator is only accurate to 1 mV, then you won't be able to tell if any inaccuracy that you are seeing is due to the function generator or the 6225.

[etechstan]

Seth, The buffer is our own design. Does NI have a specific cable for doing this or should I dismantle one that we already have? Where would i connect the Oscope?