output constant voltage DAQ NI PCIe-6321

[quixomniac]

Good Evening,

I am trying to output a constant voltage it DAQassist of 10V.
However, as far as I can tell, whenever I want to generate signals, it will only output waves.

I am relatively new to labview and cannot yet fully use the individual driver modules for DAQmx.
Currently I am using an external voltage source to power a load cell which I am trying to get signals out from LabVIEW.
I was somewhat succesful, but a cleaner set up would be to get an output from the DAQ to fully integrate the system
I know that traditionally you use a set up recommended by NI such as this:
http://www.ni.com/white-paper/7138/en/

However, I am working with existing hardward on a limited budget.
I have a DAQ NI PCIe-6321, with a BNC 20990A connector block.
So I am working with BNC connectors.

[johnsold]

The DAQ Assistant creates DAQmx code internally. Open the front panel of your DAQ Assistant (click OK on dialog), then open the block diagram. You can see how they do it. At that point you should be able to modify the AO Write for a single value rather than a waveform. You may also need to change some timing settings.

Next, I doubt the 6321 can drive a load cell. The 6341 AO lines have a 5 mA limit. A standard 350 ohm load cell requires ~30 mA. You will need some kind of buffer amplifier.

Lynn

[quixomniac]

I found the AO Write, the small icon with a pencil and wave.
Underneath it is a small list of which you can select different things,
so I changed it to 1D DBL, I do not know if that is what you meant.

So it wouldnt be worth pursuing, as in it wouldn't be possible to output 10V out of the 6321 without an amplifier?
The load cell I am working with is 700 ohm and we are driving it currently with a voltage source.
We are also having trouble filtering the signal.
The output of the load cell which we put as an input to the 6321 range from 0.8 mV to 20 mV.
Is there another limit I must be aware of when trying to input voltages to the 6321?

[johnsold · Accepted solution]

Change the AO Write to Single Channel >> Single Sample >> DBL. That should give you one value each time you call the write VI. Also change the data input to a simple numeric control. Maybe call it Excitation Voltage or something like that? Then you cna probably remove the DAQmx Timing VIs and the Regen property node as those will no longer be relevant.

A 700 ohm load cell will draw 10 V/700 ohms = 14.3 mA, which is still well above the 5 mA limit of the AO lines.

A load cell is wired like a bridge. It is the equivalent of two voltage dividers. With no load on the cell the two output lines have equal voltages: 5.00 V. When you have a 20 mV output from the load cell the voltages on the two output lines will be 4.990 V and 5.010 V (or 5.000 V and 5.020 V, depending on the internal configuration). So your 6321 must be configured to to accept 5.020 V (plus a little for noise, errors and overload conditions). So you will need to set it to the 10 V range. On the 6341 (I happen to have the specifications handy - the 6321 is probably very similar) that means a +/-10 V range. That results in a theoretical resolution of 0.3 mV. For a 20 mV full scale output from the load cell this is equivalent to 6-bit (six, not 16) resolution.

Two possible ways to improve that: 1. Use a preamplifier or signal conditioner to amplify the 20 mV to several volts. These usually have differential amplifiers to reject the 5 V common mode voltage. They may also have hardware filters.

2. Use a split excitation supply so that the load cell is excited by +5 V and -5 V. That puts the common mode voltage at ~0 V and you can use the +/-0.2 V range on the AI lines.  The resolution is then about 6 uV and the effective resolution on your signal is about 11 bits.

Lynn

[quixomniac]

Thank you for your help,
It seems I was overcomplicating it.
If I choose Single sample within the NI DAQ assist, it accomplishes what you mentioned.
Which works.
You then put it inside a while look and will constantly output the voltage you have wired to it.

I will look into amplifying the signal.
Running it thru a multimeter showed me that the signal itself was not noisy.

[johnsold]

Multimeters typically do some averaging internally so the noise could be higher than what you see on a multimeter.  However, many systems are fairly quiet.

Lynn