NI 9215 COM resistors cancel signal

[1984]

I use 2 channels of an NI 9215. Ch0 is connected to a sensor (expected signal: 0-1.5V), Ch1 measures a voltage drop on a very small (1-3mOhms, expected signal 0-3V) resistor. After collecting samples I saw excessive 60Hz noise on signals. By using a scope it became clear that the noise is added by the cDAQ system itself because I see no noise if I remove the wires from the connector of the NI 9215 and measuring the signals with a scope directly on the wires.

I am not sure how I suppose to handle this situation but what I have done is that I have added 100k resistors between the COM - Ch0+ and COM - Ch0- according to: https://www.ni.com/en/shop/data-acquisition/measurement-fundamentals/field-wiring-and-noise-consider...  (see figure12 b)

As a result the Ch0 became 100% noise free. Actually its such noise free that it rather looks like a signal smoothed/filtered. In the other hand though on Ch1I have constant 0V, so it seems adding the resistors to ch0 pretty much cancelled the signal on Ch1. If I remove the resistors from Ch0 the noise returns as well as the signal on Ch1

I have tried using 10k and 1M resistors as well, the result is the same.

I am wondering what could be the reason for this and how can I get rid of the problem.

Thanks!

[johnsold]

I suspect that you have a ground loop. It appears that you have ~1000 A flowing in the circuit measured by Ch1. If the reference for the sensor connected to Ch0 is not tied to the same potential as the low side of the shunt resistor, you could have a substantial difference in the common mode signals for the two channels. According to the specifications for the NI 9215, all input signals must remain within 10.2 V of common.  A few milliohms between the low side of the shunt and the reference point for the Ch0 sensor could result in violation of that specification.

Can you post a schematic diagram of the entire setup, particularly showing all ground/reference/common connections?

Lynn

[1984]

The circuit is simply a battery and a motor. You are right, the current can go up to ~1kA momentarily.

On Ch0 the current is measured with a hall effect sensor which provides 1V for every 1000A. The sensor is driven by a +-15V signal. (so terminals on the sensor are: +15V, 0V, -15V and provides its outputs on V+ and V- terminals)

On Ch1 I try to measure wire resistance between the battery and the motor.

I am not sure but I hope I have answered your question.

Let me know your thoughts.

[1984]

more info:

The power supply of the cDAQ chassis and power supply of the hall effect sensor are powered from a different power strip. The ground terminal screw of the cDAQ chassis is connected to the stand of the motor itself.

[johnsold]

Where is COM on the NI 9215 connected to the  battery/motor circuit? Is the battery/motor circuit connected to the chassis ground and the motor stand? Where is the 0 V point of the power supply to the sensor connected to the NI 9215 and the motor circuit?

Note that since the battery wire is part of what you are measuring, the exact connection point (which end of the wire) makes a difference.

Lynn

[1984]

Connecting all the power supplies to the same power cord made a huge difference in signal quality, so I definitely had a ground loop. Kudos for that! I have verified the measurements and the calculations and everything adds up now. I moved forward and extended the circuit with a second motor connected parallely. (I also have removed the resistors from the NI 9215)

In this case one of the current signal was quite noisy, but I have connected the COMs of both of the power supplies driving my hall effect sensors and the signal looks very good now. Great progress.

I have the feeling that something still needs to be done, since the calculated results are still quite off from the expected. I need to triple check this, but I assume that the math is good and the expectations are legit, so something has to be wrong with the signals.

The COM of the NI 9215 is not connected to anywhere now, neither the chassis ground. I am quite hopeless here, I am not sure what should I do with these wires. Do you have a recommendation?

Thanks in advance.

[johnsold]

The COM of the NI 9215 definitely needs to be connected to something. The instantaneous voltages on the AI lines may not exceed +/-10.2 V from common.

Probably it should be connected to the COM on the power supplies to the Hall devices and also (possibly through a resistor) to the negative terminal of the battery.

Lynn

[1984]

Now I see what the COM does. I have wired up the circuit as you have suggested but I see no changes in my results. All of the signals I got are very clean and noise free. So maybe its not the signals anymore. I believe this must be one of those annoying issues when the solution is just right in the front of your eyes, but you dont see it.

No matter what, kudos for the ground loop tip. Unrelated to the this issue I was assigned to another project today, which probably will take couple weeks, and we leave this one pending. I may come back later to this topic with some other questions.

Thanks for your enthusiastic support.

[johnsold]

I am glad you made some progress.  Maybe when you get back to this project, a different perspective will point you toward the solution to the remaining issues.

Lynn

[1984]

Lynn,

just to give you a feedback: I have tried several grounding methods without any success, but then we bought an NI 9215 with BNC which has an internal ground reference. That has resolved our problem.

Thanks,

i-