Significant Noise on 9213

[mtpettit]

Hello,

I recently inherited a LabVIEW based data acquisition system that among other modules, uses an NI 9213 TC module in a 9181 chassis. This module is used in an industrial environment to record temperature data from some ovens, both J and K TCs. The system worked well for many months until I recently added some new TCs to the system. Unlike the previously installed ones that worked fine, these TCs are not permanently installed in an oven. For the new thermocouples, from the enclosure box containing the chassis, TC wire runs through its own conduit to a location near the furnace where I can plug in some TCs to be used in the ovens as survey thermocouples. These new TCs are ungrounded K TCs.

My problem is that whenever one of these survey TCs is an oven and the temperature gets above ~650 C, I begin to see huge fluctuations in the incoming signal. Sampling the signal in NI MAX shows the signal jumping from what I believe to be the actual temperature of somewhere between 600 and 700 to anywhere all the way down to 200C. I originally believed this to be caused by the higher current draw of the furnace at higher temperatures inducing significant noise into the system. However this problem will sometimes still exist even if I shut the oven down and watch the signal while it’s cooling down. This problem also induces noise for every other channel on the module. When I sample the signal from the built in TCs from either the active or inactive oven, I see these same fluctuations. Some of these inactive ovens are 30+ feet away and the TC wire travels in separate conduit until they all meet in the enclosure box.

I bought some twisted and shielded TC wire too but that hasn’t helped the issue. I still believe there to be some issue with grounding, but I haven't been able to identify anything yet. Thank you for any help at addressing this issue!

[stockson]

Could you maybe draw a schematic of your system? What is the distance between the TCs and the measurement system?

[mtpettit]

I've attached an image with a rough scheamtic of the system. The TCs wires are about 15 feet long with the problematic oven. In the image the black wires are the survey TCs that are free hanging next to the machine. They're always connected to the module, but they're not always in an oven. If none of them are in an oven, I have no problems with noise. It's only when one or more TCs are in oven and above ~600C that noise is seen on that channel and potentially the other channels on the module.  

[Henrik_Volkers]

How is your common COM connected?

Isolation material behave sometimes funny when it's getting hot. All isolations migth be affected... means it doesn't have to be the ones of the new TCs 🙂

How does the noise look like?

Spec:

Each channel passes through a differential filter and then is multiplexed and sampled by a 24-

bit analog-to-digital converter (ADC). The channels share a common ground, COM, that is

isolated from other modules in the system.

[mtpettit]

There's currently nothing plugged into the COM port. I had been reading the manual on that section, and was a little confused. Should I have something plugged into the port?

I can screenshot the view from MAX the next time the system is running, but the data will read at what I believe is the right temp for a few seconds and then drop down to 300C and stay there for a few seconds (+/- 100 degrees) then jump back up. I can answer that question better the next I can check the system.

[johnsold]

COM on a DAQ board is not a COMmunication port. It is a connection to a COMmon reference potential, which might be considered a "signal ground," although it may or may not be connected to actual ground.

Any voltage measurement is the measurement of the difference of potential between two points.  For thermocouple measurements the two points are typically the two wires of the thermocouple. BUT, the internal circuitry of the measurement system is powered from a power supply which probably has one terminal connected to that COMmon point. If the voltage on either (or both) of the thermocouple wires is more than some limiting value away from that COMmon point, the circuitry will not work correctly and may even be damaged. As an example suppose that the Type J thermocouple is at 650 C. Then the voltage difference between the two thermocouple wires is about 36 mV. The voltage measurement range for the 9213 is +/-78 mV. The common mode voltage range is +/-1.2 V to COM and +/-250 V to earth ground. (See pages 18-19 of the NI 9213 Operating Instructions and Specifications). NI does not indicate what will happen if the specification is violated but the kinds of effects you report are not unlike those I have seen in other out-of-range devices.

What may be happening is that one or more of the thermocouples is picking up a signal from an oven (probably by induction) and that signal causes the voltage to exceed the 1.2 V to COM specification.  If the thermocouples on one oven are more than 2.4 V from the thermocouples on another oven, then  it is certain that at least one will violate the common mode specification. With the separation in space you described it is not at all difficult to get voltage differences of this magnitude.

Grounded thermocouples may help but you might just get large currents flowing through the grounds.  The same may occur with sheathed but ungrounded thermocouples. If you have a handheld multimeter, try measuring the voltages between various thermocouple inputs and COM. Check both AC and DC. 

Lynn