Electromagnetic Interference (EMI) on LM45 Sensors: Which terminal configuration should I use in DAQmx??

[antonio.matar]

Hello there folks,

I am using LM45 Sensors from National Semiconductors to measure the temperature inside a measurement chamber.  These sensors are a charm because they can be powered with my DAQmx NI-USB-6212 Card, directly from the +5V and DGND pins of the digital port.  Then I connect the Vout pin from my sensor to an Analog Input of the acquisition card and voilà, I'm sensing temperature.  The sensors are very user friendly because they provide the temperature directly in mV.  Say, 1V = 100ºC, 250mV=25ºC and so on. 

Anyways, these thermometers are only a minimum part of a whole system which I'm working with.  The system includes nanopositioners (voice coils, motors, etc) which use very high currents (up to 6A) to operate, and when connected and working at the same time as my thermometers, the motors induce noise (Electromagnetic Interference (EMI)???) and I sense huge peaks on the temperature which are not real.

I tried to solve this, by doing a temperature measurement every 2 seconds (Samples to Read=20K, Rate=10K).  I made an average of all the data acquired during 2 seconds, and could somehow filter the noise peaks.  I also connected the thermometers to the NI Card with coaxial cables with a connection to the earth ground to reduce the EMI.  So I solved the problem temporarily until I realized that, for the WHOLE program (apart from the temperature sensing), it was useless to operate the program at such low acquisition rates!!! 😞

Now, I need to operate the DAQ card faster, and the peaks returned!  I've tried everything, even powering the sensors via an external DC Source, etc, etc.  I wonder which "terminal configuration" should I use?? (Differential, RSE, NRSE???)  How should I connect the "digital ground" and the "earth ground"??

I know that maybe this issue has an easy answer, but I haven't come up yet with one.  Has anyone had similar experiences with EMI Noise or grounding?? I need to solve this issue soon, otherwise I'll have to change the sensors to some other less EMI sensitive mode, and maybe even my whole PID Temperature control loop! :S

Thanks for your replies,

Antonio

[johnsold]

Antonio,

Welcome to the real world!

First, you should be thinking in terms of the entire system, not just how to reduce interference to the temperature measurement.

Second, it is always better to reduce or eliminate interference at the source rather than at the receiver, if possible.

Next, it is useful to determine the coupling between the interfering source and the monitoring system being disturbed.  Is the signal radiated? Inductively or capacitively coupled? Ground loops?  Knowing this gives clues as to possible solutions.

How large are the EMI signals at the input to your DAQ card?  Measured by an independent instrument, not the DAQ card.  What waveform(s)?

How is each piece of the whole system grounded?

Lynn 

[ajmartin]

I have  had somewhat similar problems with the input of an ADC on a microntroller, motor switched on ADC goes crazy. Solved with some well placed .o1Uf caps at the ADC input pins on the micro.

Alan

[antonio.matar]

Hi Lynn,

I just did an online course regarding EMI and you're right, It's always better to eliminate the interference at the source.  The problem is that I haven't found the exact point of the source yet:

- I have a 6A DC source which powers the drives...then the drives power up the nanopositioner motors.   Between the drives and the motors there's a long (>5m), thick cable with data and power lines.  The cable is rolled up...so one theory i have is that it makes like a solenoid which radiates EM waves into the measurement chamber.

- Other theory I have, is that the voice coil itself (motor) induces the EMI, if it is so, it's impossible to electromagnetically isolate the motors, because they are moving in the X and Y axes the whole time.

Regarding the measurement of EMI waves, I don't even know the name of the instrument used for that purposes 😛

I'm about to read the NI's document: "Field Wiring and Noise Considerations for Analog Signals", and double check if I have ground issues.

If I come up with a solution I'll share it, otherwise I'll post bak again describing how the whole system is grounded.

Antonio