How to use Programmable filter for PCI 6280

[Ninad]

I am new to NI hardware and using it fot the first time.

The NI hardware purchased is 

PCI 6280, 

PCI 6528, 

RTSI bus cable,

Shielded Connector Block for PCI 6280

Connector Block for PCI 6528

I want to measure a voltage with max. of +16 V DC.

As I am a first time user of NI I have some questions,

pl. excuse me for any lack of knowledge

1. NI can't measure voltages greater than +10V so I am using a potential divider

which gives o/p as 8 V DC when 16V DC i/p is applied (50% scaling).

Is it correct ?

2. The input is connected at ai0 and ai gnd. pin 68 and pin 67 and RSE is selected

I am using the test panel of Measurement and Automation Explorer.

Is it correct ?

3. I am measuring the DC voltage and I need a precision voltage measurement and I

selected 18 bit resolution card only for this purpose. Now at present what is happening is

Measurement and Automation Explorer shows voltage as 6.9946V when

14V DC input is applied (this is OK as I am using 50% scaling by potential divider as mentioned

above) but last two digits always flicker (i.e 6.99 is constant but last two digits 46 always flicker).

I am using a DC battery source so there should be no question of noise from the input source.

The length of wires is less that 1 feet and that too they are shielded.

When I look at the graph, there is a noise. This noise remains even if I short terminals 68 and 67.

I suppose this is inherent noise of NI hardware this noise measures 0.0009V to 0.0015V

even when input is shorted.

How to solve this problem ?

4. There is a programmable filter, but how to use it ? How to enable filter in Measurement and

automation explorer ? secondly the cutoff frequency is 40 KHz. When

I searched thru. the docimentation on internet this is fixed frequency low pass filter. I suppose I need

a filter with cut-off frequency of about 04 Hz as I am measuring DC voltages.

I am not using LabView I am using Visual Basic 6.0

Can you guide me on this ?

[Raajit_Lall]

Hello Ninad,

Here are the answers to your questions

1. NI can't measure voltages greater than +10V so I am using a potential divider

Which gives o/p as 8 V DC when 16V DC i/p is applied (50% scaling).

This should be working fine. As long as the voltage that is being fed into the card is below a max of 10V you should not be having a problem. The noise that you are getting could be because of this setup though.

2. The input is connected at ai0 and ai gnd. pin 68 and pin 67 and RSE is selected

I am using the test panel of Measurement and Automation Explorer.

This is fine.

3. I am measuring the DC voltage and I need a precision voltage measurement and I

selected 18 bit resolution card only for this purpose.

The absolute accuracy of this card is +/- 1.0060mV. You can find this information using our accuracy calculator. As I mentioned above the noise that you are seeing coule be because of the voltage divider that you have setup. If you plug in a direct voltage, which is lesser than 10V do you still see the noise?

Here is another article that talks about noise considerations

4. As far as the programmable filter is concerned, it is not possible to configure is using Measurement and Automation Explorer. This is something that needs to be selected programmatically, using LabVIEW or some other programming environment.

Since you are using Visual Basic, here is a command that will set the property for enabling the filter on the channel that you select. This is a fixed filter at 40Hz and cannot be changes. You can only enable or disable it.

int32 __CFUNC DAQmxSetAILowpassEnable(TaskHandle taskHandle, const char channel[], bool32 data);

Best Regards,

Raajit .

[Ninad]

Dear Mr Raajit,

First of all my sincere thanks to you for a reply which has increased my confidence.

My basic qurry which I have mentioned earlier is

"This noise remains even if I short terminals 68 and 67.

I suppose this is inherent noise of NI hardware this noise measures 0.0009V to 0.0015V

even when input is shorted."

I am using shielded connector block recommened by NI.

My observations are;

1. Even if I do not connect a wire to terminals 68 and 67.

I am getting a the noise which measures 0.0009V to 0.0015V.

2. This remains at same value i.e 0.0009V to 0.0015V even when input is shorted.

i.e if you connect a wire between pin 68 and 67.

3. I have observed even that closing the cover of shielded connector block has no change

on noise.

All other channels i.e. ai1 ai2 etc have same noise value. If we look at the graph it is noise only

I suppose it is inherent noise of NI hardware.

I have also mentioned in my earlier querry that even when I use DC battery source (1.5V pencil cell)

this noise adds and voltage measurement always fluctuates. 1.5009 to 1.5016

I feel that Measurement and Automation Explorer should have some GUI to enable or disable

40Khz lowpass filter so that it becomes more convinient for users like me. If it's already there

then pl. inform me the steps to do this.

I still need your guidance so please help.

Regards

Ninad 

[Raajit_Lall]

Hello Ninad,

Unfortunately there is no direct method of using the programmable filter in Measurement and Automation Explorer (MAX). You can just create a task in MAX and export that into a programming environment and add the filter there.

This is pretty easily achievable. Here is how you can achieve that if you are using LabVIEW.

1.       Right click on Data Neighborhood in MAX and select create new -> DAQmx task. Click on next.

2.      Select the appropriate type of measurement that you want to take.

3.      Select the appropriate device and physical channel.

4.      Name the task and select finish.

5.      Select the appropriate settings in the NI-DAQmx settings window.

6.      Click on save on the top left corner.

7.      If you now go to LabVIEW and place a DAQmx Task Name on the block diagram, you will be able to select the task that you created in MAX.

8.      After selecting the task, you can create code out of the task by right clicking on it and selecting Generate Code -> Configuration and Example.

9.      This will create LabVIEW code for you out of the task that you created in MAX. You can now add the property node for the programmable filter in the LabVIEW code and use it.

10.   In LabVIEW, this filter is enabled by creating a channel property node and setting the property to Analog Input>>General Properties>>Filter>>Analog Lowpass>>Enable, and set it to True.

The noise that you are achieving is less than 1 mV, which is lesser than the accuracy of that card; therefore it is probably inherent noise.

Regards,

Raajit Lall

Applications Engineer

National Instruments

[nola]

The noise may be due to the RSE input configuration. The noise may be reduced by setting up a differential measurement. This can be done by setting up a differential divider, referencing the center of the divider to AIGND and measuring the signal differentially (either with either differential mode or NRSE mode).

This is illustrated in the attached image.

Additional DC measurement improvements can be made by averaging. If the noise is due to 60 or 50Hz line nose then averaging over an integer number of line cycles will help filter out the unwanted interference.

Adding a source resistance to the input by adding a resistive divider will add noise proportional to the square root of the equivalent source resistance and the square root of the bandwidth. For reference 1k ohm adds 4nV in a 1Hz bandwidth.

Adding resistance to the input also reduces how fast the channels can be scanned, the larger the resistance the slower the channels can be scanned. Of course if the application is just sampling from a single channel then scanning doesn't apply.

Loading of the input source by the resistive divider should also be considered. Both in terms of the gain error due to the source resistance and divider combination and the current draw from the load.

Message Edited by nola on 01-30-2007 08:30 AM

[Ninad]

Dear Mr Nola,

My sincere thanks to you for a valuable reply.

I have studied the .jpg file attached to your reply.

I am building a test jig for electronic circuit and for this application I need to measure DC voltage at 11 points in an electronic circuit. So I selected 18 bit, 16 AI card. Now as per your recommendation

if I am using AI0 and AI8 for differential measurement then I feel I need 22 channels for such

configuration. i.e. 2 channels for each differential measurement. Am I correct ? I am a first time user of NI hardware. pl. excuse me for any lack of understanding. If this is correct then my NI hadware would be insufficient.

Right now, I have increased the no. of samples to 1000 and I am averaging it.

What is the actual use of AISENSE terminal. I could not understand.

From your reply "

Adding a source resistance to the input by adding a resistive divider will add noise proportional to the square root of the equivalent source resistance and the square root of the bandwidth. For reference 1k ohm adds 4nV in a 1Hz bandwidth. "

I could not understand this. Can u pl. expalin this in further details.

Right now. my potentail divider values are

R1 : 10K

R2: 9K1

R3: 560ohm

R4: 1K, 10 turn preset

R1, R2, R3 and R4 are in series.

Input 16V DC is applied between R1 and R4,

I tuned R4 in such a way that I get 8V DC across R2, R3 and R4. This 8V DC is

applied to AI0 channed (pin 68 + and pin 67 -)

Is this OK or can you guide some other arrangement.

Regards and Thanks again for your valuable reply

Ninad

[Raajit_Lall]

Hello Ninad,

You are right, if you are using a differential configuration, you would indeed need 2 channels per measurement. This would mean that you would need more channels than in Single-Ended-Ground Referenced (RSE) or Single-Ended-Nonreferenced (NRSE) mode.

The AISENSE pin is used for the NRSE mode. All these signal source types are discussed more in details on this website.

Field Wiring and Noise Considerations for Analog Signals

Nola, could you provide further details on Ninad's third question please?

Regards,

Raajit Lall

Applications Engineer

National Instruments