Multifunction DAQ used as DVM

[johnoc]

Hi Ho!

Has anyone used a Multifunction DAQ card as a Digital Volt Meter? I am using a PCI-6254 and LabWindows and trying to measure Va and Vb from an RVDT. From those measurements I calculate gain to determine angular position. The gain figure needs to be accurate out to 4 places (0.XXXX). My measurements with the card are close but still not what I get with a couple of calibrated fluke DMM's. The measurement with the card seem to bounce around a lot as compared to the flukes even it I change the update rate of the card to match the DMM's. At 16 bits, I would think that I would be able to get closer  than I am.

I suppose that there could be noise causing problems. Any suggestions for software filters to settle the measuremnts down? Has any one even tried to read RVDT's with their own signal conditioning?

Any help would be appreciated, since I am having to profile 22 RVDT's in one system. We have a final solution but the solution will not be available for another couple of months. I am trying to provide an intermediate solution.

Regards-

John 

[Robert_F]

Hello John,

The NI-6254 is capable of acquiring the data from your RVDTs.  You would probably want to use all 16 analog differential inputs on the NI-6254.  If all the RVDTs can be referenced to the same ground, then you could measure 32 analog inputs under the single ended mode.  Differential mode is going to use two analog input channels to acquire data.  They are paired up according to the following table.

A0 – A8                       A16 – A24

A1 – A9                       A17 – A25

A2 – A10                     A18 – A26

A3 – A11                     A19 – A27

A4 – A12                     A20 – A28

A5 – A13                     A21 – A29

A6 – A14                     A22 – A30

A7 – A15                     A23 – A31

The NI-6254 is going to have a range accuracy of 1.92mV with a Range of +10V to -10V.  If you are getting a lot of noise with your signals connected in the single ended mode, you might want to try differential mode.  Here is a great document that describes a few basic concepts on eliminating noise from your measurements.  I would use the schematics in Table 1 of this document to connect a few biasing resistors to your circuit.

[johnoc]

Hi Rob,

Thanks for you response. I decided to go a different way for the time being. The 6254 will be used in our final solution but it is just not accurate enough when I am measuring the two ac voltages calculating the rms and then the gain on the RVDT. I tried adding averaging and my numbers improved but we still were not where we needed to be. For this type of measurement we may have to move to an 18-bit card or even higher if they're available in the future. An 18 bit card would probably be good in the +/- 5 volt range. As I said we need the accurracy on our gain out to the ten thousandths place.

In our final solution, converting the two AC signals to a singal DC signal and calulating gain. The 6254 will be more than ample enough for the job.

Thanks for your help!

Regards-

[ChrisR.]

I know nothing about RVDT's, but I'd be very surprised if you can do much better than the 6254 for AC measurement, unless the signal frequency is outside the range of the 6254, i.e. greater than about 500kHz. Computing rms requires good windowing and computation technique, but it's much more accurate than conventional rms-to-DC converters. It is the technique used by NI's FlexDMM products. How are you planning to convert the AC to DC? Or do I misunderstand you?

[johnoc]

Chris

Analog devices makes a chip that is specifcally designed to provide the the exitation signals as well as demodulate the outputs of RVDT's and LVDT's. We are constructing a fixture that will allow us to handle all 22 RVDT's at one time using these chips. We will use the 6254's to measure the DC outputs of this fixture.

Regards-