crosstalk between pxi6120 AI Channel

[FrankyCam]

I have the following application:

  -a pcb card with four amplifier for extensiometric signal, with floating differential output -5/+5V

  -a PXI6120 pci card with 4 analogue psudodifferential input

When i connect only one of the four extensiometric amplified floating signal to one of the AI card, all works fine. Instead, when i connect two or more extensiometric signal to PXI card, crosstalk between channel occurs.

I just made the following test

-measuring with tester pcb card output and all works fine

-acquiring signal with a USB NI6008 card and all works fine

-putting a resitor of 20KOHM between AI(-) and AIGND, for each channel; corosstalk occurs again

-disconnect each not used cable form PXI channels

Any suggestion?

Francesco

[Norbert_B]

First of all, this question does not refer to LV. So you should have posted it in the DAQ forum.

Second, cross talk usually occurs due to high impedent signal sources. You should consider to use impedance converter to lower input impedance for the data acquisition.

Norbert 

[FrankyCam]

Thank you for the quickly answer, and excuse me for the wrong session. How can I move to daq forum this discussion?

Another question, what is an impedance converter?

francesco

[Norbert_B]

An impedance converter changes impedance of signal lines. There are different kinds of impedance converter, most often they are used in audio technology. Termination resistance for different applications is usually 50/75/120 Ohms (audio, video, ...).

For general voltage acquisition, the rule is that the source should have 0 Ohms impedance, the sink infinite.

For actual measurement systems, the sink (DAQ device) usually has 1..10MOhms, sometimes more. So it depends on the source impedance if the measurement is

a) accurate

b) crosstalk can occur.

If you have a signal source with 1MOhm and the DAQ device has 1MOhm, your signal will be split half between the signal source' impedance and the DAQ device. You will measure 50% of the real sources signal (100% measurement error!).

If the source has 1kOhm, the signal will still split, but at a ratio of 1:1000. So your measurement error is still 0.1%.

Missaligned signal source impedance is for multiplexing devices a problem since the abundance or absence of electrons has to be balanced out over the signal source.

Since multiplexing switches channels as fast as possible, this balancing process is probably not finished, hence you will see cross talk. 

Cross talk can be seen starting with 100Ohms and higher signal source impedance (rising %). So you will need either

- to slow down multiplexing and DAQ (possible, but must be implemented)

- adopt signal source impedance to 100Ohms or less.

Norbert 

[Norbert_B]

Well, reading seems not to be my favorite today 😞

You are using a S-series device, so there is no multiplexer on the channels.... So my previous post is correct, but will not apply to your issue.

Re-reading your initial post, i would say you have a grounding issue. If your ground reference is supplied with noise (from any source), it will affect all channels if you don't measure differential. Pseudodifferential acquisition does refer the signal to the ground reference of the DAQ device as seen here. So noise introduced by one or more channels will affect the ground reference of the DAQ device, creating errors on ALL channels alike.

Solutions are:

- change to differential measurements (not sure if the 6120 can do this, had to read the specs)

- use filters, possibly in combination with isolation

Norbert 

[FrankyCam]

Another question. I suppose that the pxi6120 is four independent ADC without multiplexing. I'm wrong?

Thank you very very much for your explanation!

Francesco

[FrankyCam]

ok for not multiplexer 🙂

so, why this problem not occurs when i use usb6008?

[Norbert_B]

---

@FrankyCam wrote:

[..]

so, why this problem not occurs when i use usb6008?

---

I assume that you are using the 6008 with four differential channels, not 8 single ended....

Norbert 

[FrankyCam]

The broblem seems to be solved with a particular ground arrangement. Thank you very very much for your help and he explanations.

Francesco