Signal distortion as amplitude increases

[Oley_Mich]

Hello community!

I'm using several NI-9232 modules to monitor dynamic outputs from an amplifier.  Amplifier channels are loaded with 4 ohm resistors.  So effectively, I'm measuring across the resistors.

I am also measuring the channels with an Oscope.

At the low end of output, I get good signals on both.  Signals are Sine waves with a frequency of 1 kHz.  The NI-9232 modules are configured for +/- 30 V and running at 102.4 kS/S/ch.

As the input signal is increased, the NI-9232 channels begin to distort well before our target output.  The scope is showing good signal, but the NI DAQ channel is distorted beyond effective use.

I have attached 4 files as reference.

This is becoming a very time sensitive issue, and I'm out of ideas to fix.

Am I doing something wrong?

Thank you for your time!

[this_one_time_at_vancamp]

Oley_Mich,

Does this behavior occur on every channel of every module?

Out of curiosity, does reducing the sample rate or signal frequency effect the shape?

VanCamp

[Oley_Mich]

Yes it does.

And as I slow the speed down, nothing changes.

I didn't go too slow as much more would ruin the acquisition purpose.

[BertMcMahan]

I'm not familiar with those modules in particular, and you didn't upload any code so this is just a guess. Make sure your acquisition is grounded properly and configured correctly. For example, many cards offer a selection of Differential mode vs. Single Ended mode- make sure your acquisition is set right.

Also make sure you're not in Current/IEPE mode or anything like that. The manual here http://www.ni.com/pdf/manuals/373932a_02.pdf shows there are a bunch of ways this thing can run (another example, AC vs DC mode).

Double check your settings. Also, double check that the scope is connected the same way (with the same grounding) as your DAQ devices. Are you measuring the scope signals and the DAQ signals simultaneously, or are you disconnecting the DAQ to attach the scope? Scope probes are typically grounded, whereas this card (appears) to be able to do differential measurements.

[Oley_Mich]

The chassis is grounded, and these modules are pseudodifferential only.

I've never touched any of the Current/IEPE mode settings and all channels are set for DC.

The scope is using a differential probe to measure same channel at the same time.

Very perplexing at this point.

[BertMcMahan]

Is your amplifier output floating or is it ground referenced?

Have you tried grounding the AI- side of things?

Sorry for just shooting in the dark here, I haven't used those modules. Just guessing 😕

[Larry_Morgan]

Oley_Mich,

The 9232 is isolated from Earth ground, but the channels are not isolated from each other (bank isolated). In each AI- input, there is a 50ohm resistor to the module's isolated common. If, by chance, your system is creating a voltage potential between the AI- terminals of different channels on the same module, this could begin to create some undesirable effects if the voltage potential difference is too high. First, it will attempt to flow current between the AI- inputs, and eventually, will activate protection circuitry designed to protect the 50ohm resistors so that they are not damaged. If the protection is being activated, I can imagine it would create some unexpected measurements. 

I mention this because generally the module performs very straightforward, but this might be a nuance that is contributing to what you see. Your oscilloscope may show a different result because it is doing a differential measurement. 

[BertMcMahan]

If this is the case you should be able to disconnect all but one channel and get good measurements (I think- is that right Larry?)

[Henrik_Volkers]

The 9232 spec tells:

Channel-to-channel common-mode voltage range
(AI– to AI–) ......................................±1 V max

So if your amplifier has a floating output (or an internal xxxmOhm current monitor shunt to GND, or...) you can run into problems on multiple inputs.

As a Test

measure only one channel

measure (one channel or scope diff)  between two  AI- amp outputs while pop-up the volume and have a look if you reach the 1V limit ...