Compensate for accelerometer signal decay due to AC coupling with NI9230

[ManfredAmann]

Hello LabView experts:

I'm trying to acquire the acceleration signal of a test sled using a PCB triaxial IEPE accelerometer (Model 356A15 - specsheet attached) and the NI9230 card mounted in a cDaq-9132. The constant current exitation is provided by the NI9230 card (internal) and the coupling is set to AC. This configuration doesn't allow for DC coupling.

The attached picture shows the measured acceleration of the test sled. However, due to the signal decay caused by using AC coupling (as shown in the second picture), the acquired acceleration signal is destorted. Do you have any recommondation on how to correct the signal in software for the decay caused by using AC coupling?

Any advice is greatly appreciated.

With best regards,

Manfred Amann 

[billko]

I'm not familiar with the term "signal decay"?

[ManfredAmann]

Hi Bill:

Thanks for your reply. By "signal decay" I'm refering to the exponential voltage decrease that can be observed when utilizing AC coupling. Without the decay, the signal voltage from the ICP transducer would remain constant as long as the load (in this case acceleration) remained constant. However, because of the AC coupling, the voltage decays exponentally and approaches zero over time even if there's no change in load - please see the second chart of my original post for more details. I'm trying to find a way to correct the acquired acceleration signal from our test sled for this decay in software.

Do you have any suggestions on how I might be able to accomplish this?

Thanks again,

Manfred Amann

[mcduff]

You have AC coupling which can be thought of as a high pass filter, that is, you are blocking DC. Thus with AC coupling it is impossible to see a static signal that is non-zero. (Get a scope and put a square wave into it with AC Coupling. Lower the frequency of the square wave and you will get what looks like a derivative signal.)

Mathematically, your signal is convoluted with a high-pass filter; what you need to do is deconvolve it with the AC response.

mcduff

[James_McN]

What is it that makes DC coupling not possible? The module appears to support it.

The problem is the DC has been filtered and I don't think there is any way to get it back. If you think about it if you have a flat line at 0Vac, there is no way you could work out that is actually sat at 0V, 10V or 1,000V DC.

We can see from the decay line that there is likely DC but even then we may have just added to existing DC offset.

The reason I ask about DC coupling is if there is a problem with that method it might be easier to solve!

[mcduff]

If it is an accelerometer it may need to be biased, if so, most NI modules enforce AC Coupling when biased.

mcduff

[James_McN]

Ah yes sorry - it has been a while since I have used one of those!

I think this is because a Piezo based sensor can only measure AC? Looking at that datasheet they give a 10% range only down to 1.4Hz. This page also states this and you may need a mems sensor: http://volersystems.com/v-2011/125-how-to-avoid-big-errors-using-accelerometers/