Showing results for 
Search instead for 
Did you mean: 

Reading inaccurate voltage from Omega Super MCJ

I am trying to read the voltage coming from an Omega MCJ thermocouple-to-analog voltage adapter using my usb-6008. (Shows temp as 1degC/mV) The voltages read by LV are lower than the values read by my meter off the outputs of the adapter(which is accurately displaying the temp) apparently due to periodic dips in the voltage.
Oversampling did not work in fixing the problem.  
I have tried RSE and differential modes with voltage buffers, with the same effect. Shorting the two AI pins together yielded a slightly negative voltage (2mV)
Is this a problem due to the low resolution of the DAQ since I am looking at mV scale differences? Would a filter sub-VI help out with the noise, or is it a more intrinsic problem with the DAQ?
Thanks for your help!
0 Kudos
Message 1 of 3


The first thing to keep in mind is that you are measuring a floating signal source.  In terms of signal connections, this means that you should either use RSE mode, or use bias resistors in differential mode.  Since your signal source has no ground reference, the bias resistors are necessary in differential mode to provide a return path to ground.  This is described in Field Wiring and Noise Considerations for Analog Signals.

As you alluded to, the USB-6008 may not be the best hardware to measure this millivolt signal.  If you look in the USB-6008/6009 User Guide and Specifications, on page 19 under the Analog Input Specifications, it lists the absolute accuracy as 1.53 mV when using the +/-1V range of differential mode.  When you switch to RSE mode, there is only one range (+/-10V), so the accuracy decreases to 14.7 mV. 

If you want more accurate USB hardware, you should consider either the DAQPad-6015 or the USB-9215A. 

Hope this information helps.

0 Kudos
Message 2 of 3
I found a similar problem and found that adding a small (0.1 microfarad) capacitor across the input at the A/D board (running in differential mode) improved the stability of the signal a lot.  It's still not perfect, but much better than the original measurement.
0 Kudos
Message 3 of 3