Dynamic Signal Acquisition

[j1204]

It doesn't matter  the sensitivity ，if  I  set  the sensitivity  to 10mV/g, the  amplitude will expand 10-times.

[WillPwr]

Agreed - and that's part of my point. The difference between 1.3 and 0.9 g may be considered reasonable within error of the accelerometer (depending on the accelerometer), while 13 vs 9g would seem not to be. I was just curious if you'd considered the scaling when examining the issue, and by how much your data was actually off.

Were you ever able to try the suggestions from John S? I'd also be curious about the brand/model of accelerometers you're using, if you have the information available.

Thanks, j1204!

[j1204]

We use B&K  4525-B accelerometer , this is the report of NI：

How do you reproduce the problem?

Plug customer’s PCI-4472B into PCI slot, open test panel in MAX , select AC coupled, all channel had approximately10mV offset without signal connected. Did self-calibration in MAX, the problem still existed.

Plug a good PCI-4472B, there is no offset on the same test condition.

[WillPwr]

Hello, j204!

Thank you for posting the information on your accelerometer. I'm a little confused about your last post - are you saying that using a different PCI-4472B seems to fix the issue? If so, then this is almost certainly a hardware issue, and the board should be sent in for repair. 

When submitting for repair, you will need both the part number and serial number for you board. Both are 6 to 7-digit alphanumeric numbers. Note that "4472B" is the model number.

If this is not what you mean, could you re-phrase please?

Thanks!

[j1204]

Thanks for you!  I've already sent this board to my local NI office，this is the  test report of my local NI office.

                                                                                            Test Report 

Name: PCI-4472B

PN: xxxXX

SN: XXXXX

Tester: XXXXX

Specific Symptoms:

Customer said all channel had approximately 10mV DC offset without signal connected even when AC coupled.

How do you reproduce the problem?

Plug customer’s PCI-4472B into PCI slot, open test panel in MAX , select AC coupled, all channel had approximately10mV offset without signal connected. Did self-calibration in MAX, the problem still existed.

Plug a good PCI-4472B, there is no offset on the same test condition.

What is your HW set-up?

Customer’s PCI-4472B.

PCI-4472B(proved good), PC.

What is your SW set-up?

Windows XP, MAX 5.1.

What were the environmental conditions under which the unit was operating?

It was running in an office environment at room temperature.

[Charlie_P]

I want to caution not to call a board "good" or "bad" based on an invalid test.  To reiterate and expand on a point that John S made earlier, the performance of Analog Inputs is not specified when they are left open.  In order to perform a valid test of an AI channel, something must be connected to it.

There are several options for what to connect to the input in order to perform a valid AI test.  The most common, widely accepted option is a 50 Ohm terminator.  Another option is a shorting cap.  A shorting cap is essentially the same thing as a 50 Ohm terminator but it is a dead short, instead of a power dissipating resistor.  The third, and arguably best, option is a calibrator.  These are precision instruments that output a very stable, very accurate voltage used for calibrating devices.  Generally, labs and offices do not have this device but if you have access to one, you should consider using it.

Please choose one of the methods explained above and post your setup and results to this page if you are interested in further discussion.

[j1204]

thanks for your suggest Piazza ，
We have sent this board to my local NI office，so I don't have one to try what you suggested. We will try those when they set it back.