Interference between channels using a cFP-TC-120 and CB-3 connector block

[Leigh]

Utilizing a CB-3 connector block with a cFP-TC-120 module to read millivolt sensors. Sensors are pressure sensor with an excitation of 10V. Millivolt signal ranges from 8mv to - 70mv. Have all 8 differential channels wired in two of these modules sitting in slots 3 and 4 of a 4 slot back plain. There is also 24V in the the electrical enclosure utilized by two cFP-DO-421 relay modules driving valves that open separate chambers to vacuum or vent to ambient, these are in slots 1 and 2. The cFP-TC-120's do the same thing and each is affected in the same way plus the interference is not only between channels on the same 120 module but also between modules! If a channel/sensor is under vacuum where the millivolt is around 7mv its reading is stable through the following conditional changes. "If a channel/sensor is vented to ambient, say channel 6, and channel one is already at ambient its reading may change by  0.25mv being influenced by the venting of other channel. This is compounding proportional to how many channels are vented by up to 0.5mv for every other channel which is vented. As this module is differential and each channel and module are isolated we cannot see where or how this interference is taking place. Help!

[BenDAE]

Leigh,

Thanks for that detailed description. Sounds like a great use of NI hardware. I'm sorry it hasn't been working right for you.

It does sound like your dealing with noise on your channels. I am providing the go-to reference on that here. http://zone.ni.com/devzone/cda/tut/p/id/3344

If by chance, what you are seeing appears more like ghosting (signals bleeding across channels), then below I have listed some steps to deal with that.

The effect of ghosting can be reduced through any of the following methods:

The introduction of a large value pull-down resistor will prevent the input impedance from increasing too high, and will allow a path for excess amounts of charge to ground.

Decreasing the speed at which the multiplexer switches between channels can also help prevent ghosting. This gives the multiplexer time to “bleed” some of the excess charge and therefore limit the effect of ghosting.

Grounding of intermittent channels that are not being read can also reduce the build of charge. As the multiplexer connects to the grounded input the excess charge will be dissipated. 

[Leigh · Accepted solution]

Ben,

We did some trouble shooting and isolated the sensor millivolt signal from the NI. We did find that the change in voltage was real and in the electronics not the field point. The system runs on 24V except for the sensors which have a 12v excitation. We found that the commons between these DC voltages were just that, each were shared. We separated the commons and vola! Perfection.

Leigh

[BenDAE]

Leigh,

BRAVO! I'm delighted for you and your vacuuming/venting application. Great work.