Cross-talk between signals connected to NI 9207

[KarinaTaylor]

I have a humidity/temperature sensor and a pressure transmitter connected to a NI 9207 in a cDAQ chassis.

The humidity/temperature sensor requires 24VDC, while the pressure transmitter requires 5VDC supply.

The wiring diagrams and the corresponding data sheets are attached. The power supplies share a common ground.

Each one of the instruments works perfectly fine when connected on its own to the DAQ module. However, if I first connect say the humidity/temperature sensor, and then the pressure transmitter, I notice a step change in the humidity/temperature signal when the pressure transmitter is connected. Then, if I disconnect the humidity/temperature sensor, I also notice a step change in the pressure signal.

I have tried two wiring configurations, as shown in the diagrams, without any change in the unwanted behaviour.

Any suggestions on what am I doing wrong?

[Henrik_Volkers]

The crosstalk you see is sadly typical for a lot of multiplexed input DAQs.

And even more sadly this behaviour isn't stated in the specifications of the 9207. If you look at specs from other multiplexed  DAQ devices the first diagram in the analog input section show the error ...

Problem is the relative high output impedance of your voltage out of the hum/temp sensor.

Since these values usually don't change that fast, can you use digital (rs232/Modbus) output of the device?

Other options:

-Use the current output, these are low impedance ...

-choose the lowest samplerate possible

-add an input buffer

-choose a DAQ with simoultanious sampled or buffered inputs

-another method if you use 3 analog voltage inputs: connect the temp and hum channel each to 3 successive channels , read all channels and use only the last of the three..., again, lower samplerates help.

[KarinaTaylor]

Hi Henrik,

Thank you for your advice. Would you mind clarifying for me which bit of information made you think that the humidity/temperature sensor has high output impedance?

Unfortunately it took a great deal of effort to find a sensor which meets all of our other requirements, and we had to decide the type of output at the time we ordered it, so now we're stuck with it. It's more likely that we shall purchase new pressure transmitters with 4-20mA output instead.

Greetings from Australia,

Karina

[Henrik_Volkers]

If the rated current consumtion is about 1mA and the current can increat another half mA at low loads...   I assume a high output impedance...

Another trick you can try: Get three capacitors (470nF as a first try, , ...)  and connect them across the inputs of a channel. at the DAQ.

How fast do need to sample? Once per second ?

[Henrik_Volkers]

Checking the Vaisala spec (Note to other readers: It really helps to post information!)

I found:

External loads

0 ... 1                         RL min 10 kΩ

0 ... 2.5 / 0 ... 5 / 1 ... 5 V     RL min 50 kΩ         

Your DAQ (static one channel is >1GOhm)  but only static one channel.

An interesting test would be to hook up only one channel (1G load) and add a 50k resistor and monitor the voltage drop.

[KarinaTaylor]

Hi Henrik,

Thank you for your advice. This definitely gives me something to work with.

I'll let you know how I'm getting on.

Cheers,

Karina

[KarinaTaylor]

Hi Henrik,

It turns out it wasn't ghosting after all, just the good old traditional ground loop. The measurements are now correct, after we removed the 1MΩ resistors (but still left the connection to the COM in place) from the wiring diagrams I attached to the first message in this discussion.

Thank you for your input anyway, lesson learned.

Karina