Fluctuating Amperage Reading

[Mech322]

A LabVIEW program is currently reading fluctuating amperage readings. The problem started after the installation of an Omega TX86. The program read the milliamps (4-20) of an Omega TX64 through an NI USB-6211. The program worked fine for several years and I didn't create this program. Then the Omega TX64 died and it is no longer made. It was superseded by the TX86. The settings of the TX86 are the same as the 64. This was confirmed by calling Omega. The 86 doesn't use a ground wire, which was never attached to the DAQ. The TX86 has more abilities than its predecessor but aren't needed. 

The TX86 reads the temperature accurately on its digital display. However, the readings fluctuate wildly in the LabVIEW program. A multimeter was used to read the amperage readings. On the multimeter the amperage readings are stable and rise and fall as expected. The LabVIEW program was cut up to focus on this problem.

Other things that have been tried: changing the USB cable, switching the USB-6211 with a MyDAQ, changing Ai ports, switching between amperage and voltage readings in LABVIEW, and adding a shunt resistor.

It keeps fluctuating. It should be reading a steady 6 amps at the current temperature. Any help would be appreciated in resolving this.

[santo_13]

If swapping the 6211 with myDAQ still resulted in fluctuations, then it is a possible software or electrical issue outside the DAQ.

[Mech322]

Hi Santhosh,

Thanks for replying.  I found it strange that the multimeter has a stable reading of the amperage change but the DAQs don't.  Any ideas what might cause that? I found the line also has a voltage under 5V.  

[santo_13]

The multimeter has a slow update rate, i.e., sampling rate, which averages out the fluctuations, whereas the DAQ samples at a faster rate and hence can observe events at a smaller time scale.

[AndyClegg]

Just a couple of comments that may help you track things down:

• Can the temp really vary that rapidly - 0 to 70 (degC?) in just one sample? Lets assume not and what you are observing is wrong, it is either the sw or the wiring that is at fault.
• Have you checked the output of the device is compatible with the AI, and with the type of connection you have ? For 4-20mA there are different types - two-wire, three-wire and four-wire - based on where the power is supplied to the device (from the IO module or externally). You need to make sure you device producing the 4-20mA can be wired in the correct way that is compatible to the DAQ module. It may be a straightforward two-wire connection.
• Check the connections are correct (including the wiring output form the device - dont assume it is the same coloured wired as the old model) - could the AI pin not be actually connected to anything and just left open / measuring noise.
• There may be an issue with floating power supplies if they do not share a common ground reference, but that depends on the connection type. Probably not the case here.
• Noise can be introduced - but again it looks like you have signals ranging from 4-11mA (? Im assuming as no units shown) - and that seems too big for normally connected signals, but depends if you have poor cables and close to something with big electrical interference.
• Multimeters do smooth signal variations - but it depends on the type of multimeter - so it is not just a case of slower sample rates. But do take care with how you are measuring if you are trying to connect a mutlimeter to the 4-20mA (in series?, or disconnecting the connection to the IO module?)
• To test it isn't the sw what about writting a dummy fixed value or sinusoid to the input (as close as possible to where AI read) and make sure you see the same signal coming out. In some of LV tools there is the ability to insert a "soft" test value at the AI pin (replacing the real signal) to test.
• There are plenty of places where variability can be introduced when dealing with signals - mis-matched sample rates in IO module and reading/processing in sw, ailiasing, lost data points, sw bugs, etc BUT LOOK AT WIRING FIRST.

[Mech322]

Hi Andy,

Thanks for responding.  I'll answer your points to the best of information that I have:

• I know the reading is wrong.  The multimeter was tested at the DAQ  AI port +/- and it showed a steady 6mA. The multimeter was also put in series near the TX86 and it had the same reading.
• The wiring is correct according to the diagram and Omega. Omega was called to confirm the wiring of the power and thermocouple. 
• The wiring has been checked repeatedly by me and others. All connections were traced and found to be correct.
• The TX86 doesn't use a ground wire unlike the TX64.  I attached a ground on the (-) to see if it would take away the noise but it didn't.
• The TX86 ranges from 4mA to 20mA
• Is sw for switch or short wave? Is there a reference to writing or making a test value? This DAQ is taking another 4mA-20mA signal from another device and reading it fine. I'm a mechanical engineer and a lot of electrical terms are new to me.
• I have also tried an Arduino Uno and it produced a very similar result. I'm guessing at this point the devices, except the multimeter, can't read the signal properly. It is also possible that noise is being introduced somewhere too. 

I have tried to slow down the sampling rate to read the signal in the while loop but the noise is the same. I have tried as slow as 5 seconds.  Is there another way to do this.  From the data sheet on the TX86 it claims a 1 second response time.  Would this cause the issue?

[santo_13]

You shared new info that means something.

"This DAQ is taking another 4mA-20mA signal from another device and reading it fine." - what device and how is it connected?

Since you're a mechanical engineer, short note: the inputs of 6211 are not isolated, and any noise due to one channel will affect others.

This will not show in your multimeter because it is powered by a battery and isolated from the rest of the channel noises, and the same, I believe, is true for Arduino.

[AndyClegg]

Yes if you have a second 4-20mA connection that looks like it is being read fine, a very quick test would be just to swap over which AI they are connected to - i.e. does the known working 4-20mA give the correct value in the sw when it is connected to the AI that is not working with the TX86. This will tell you if it is the sw (software) or external (wiring) that is causing the problems reading from the TX86. If the sw is not working you will at least be able to debug using a known working signal.

My guess is that it is a wiring problem, due to big random fluctuations in the current from the TX86. Did you connect a 24V DC power supply to power the 4-20mA as shown in the diagram. Even if you did, it is important to understand how that relates to the input circuitry of the AI module you have and what shoudl be connected to AI+, AI- and COM.

This forum post has an image that may help: https://forums.ni.com/t5/Multifunction-DAQ/Connecting-a-4-20mA-loop-powered-sensor-to-USB-6255/td-p/...

[Mech322]

Do you know how to isolate the or eliminate the noise?  I changed the power supply to see if it was the problem.  The power supply wasn't the problem.  It appears after all I've done noise is the problem.  

[AndyClegg]

It might help if you do a wiring schematic, showing *all* of the connections between the TX86, sensor power supply, the USB 6211  (or MyDAQ - not sure whcih one you are using). Does your IO module actually accept mA input, or is it a voltage input and you are using an external resistor to convert from mA to voltage ?

However, this forum post talks about noise and wiring for a USB 6211, and may help you work out where there may be something wrong:

https://forums.ni.com/t5/Multifunction-DAQ/USB-6211-noise-problem/td-p/2628867