Can I use a PCI-6251 to bias a sensor?

[Fascinated by Electronics]

Hi,

I am currently working on a measurement project where I need to use an accelerometer to measure vibration in one of our instruments. In our current setup, we are using a triaxial accelerometer that is powered by an I/O module made by the same company. It is a bit overkill in terms of the sensor's specs and its DAQ software when compared to our needs. Therefore, we are looking to slim down by using LabVIEW to collect the data.

Making a LV program to acquire the accelerometer data is fairly straightforward, but the more important question I have is: Can I used a PCI-6251 or USB-6221 to set the sensor's bias? I know that several LV DAQ boards have analog and digital outputs. From the sensor's spec sheet, it needs a supply voltage of 8 to 32 VDC and a supply current of 9 to 12 mADC.

I checked the specs of those two LV DAQ cards and it seems like with the analog output, they can source the proper voltage (+/- 10 V), but their current output is only +/- 5mA. Is it true that my sensor cannot be properly biased with those two aforementioned DAQ boards? Is there an easy to increase the current level up to say +10mA ?

Once I know the answers as to the hardware needed to source the correct supply current, the LV programming for it should be relatively easy. Just have a structure where the analog output of the DAQ is instructed to output DC waveform at 8-32V and 9-12mA.

Thanks!

[jschwartz]

Hi Fascinated,

You are correct for the current and voltage specs of the 6251 and 6221.  You could take an acceleration measurement with these boards if you have external current excitation. A better option would be to use a DSA device.  These devices have 24 bits of resolution, can provide current excitation and are optimized for acceleration measurements.  It sounds like you have not already purchased hardware.  If this is the case, you should take a look at the  PCI-4462 which can supply 10 mA of current excitation to your accelerometer.  

NI PCI-4462

[Fascinated by Electronics]

Thanks for the reply Jim. I will look into the NI hardware that you mentioned.

[Fascinated by Electronics]

Hi.

I just thought about a possible way to bias my accelerometer using the previously mentioned DAQ cards (PCI-6251 or USB-6221). Please excuse my limited Labview and electronics knowledge if my proposed solution is not possible due to any number of reasons (hardware, electrical theory, etc)....

Since both the 6251 and 6221, have TWO analog outputs..... can I use both outputs? That is, can I program AO 0 to output a DC waveform at 5mA and AO 1 to output a DC waveform at 5mA and then ADD those two signals and then send that signal to my accelerometer?

Gurus and electronics experts, please let me know if this is possible....

Thanks

[ChrisR.]

It may be possible.  Can you post a link to the datasheet of the transducer you're trying to use?  And are you mentioning those two DAQ products because you already have them?

Chris 

[Fascinated by Electronics]

Hi Chris, I already have the 6251 and 6221  DAQ cards. I have attached a PDF of the accelerometer we are leaning towards purchasing.

[ChrisR.]

It looks like you could make it work, although the worst-case current requirements of the sensor (12mA) is more than the guaranteed current available from two combined analog outputs (10mA).  Chances are good that the sensor won't require more than 10mA, and that the 62xx boards could deliver a little more than their rated output current if called on to do so.

In order to ensure current sharing between the two analog outputs, you'll need to put a small resistor in series with each output.  I recommend setting both to +10V and connecting each to the transducer power pin with a 100-ohm resistor.

Chris 

[Fascinated by Electronics]

Thanks for that info Chris. It looks like we are leaning towards getting a dedicated power supply for our current and future studies. We are borrowing a DC power supply right now. Things are working well and the accelerometer responds when force is applied to the test item. However, there are still some things about the voltage trace and later data manipulation that I am unsure of. I am still a novice when it comes to data acq with LV 8.6 and therefore have some more questions:

1) How do I remove noise? I haven't determined the frequency spectrum of our measurements of interest, but right now things are very noisy at baseline. At most, we see about a 2 order of magnitude increase in the voltage amplitude during our tests... So, indicated RMS at baseline is about 0.0005 V and max RMS during the test is around 0.02 V. Are there things I can do the Waveform graph display itself to lessen the effects of noise? Do I need to add some filtering into the program to remove what I assume is 60 Hz noise?

2) For some strange reason, during the test, with our LV program running, the X-axis  (which is Time in seconds) does NOT appear to be moving..... that is... the end point of the time display does not appear to be increasing......... It always displays 0.25 s on the X-axis, even though the program is still running and Y-axis data is changing.... Is there something wrong? I have the DAQ setup for continuous samples, so it shouldn't be stopping after a predetermined time............

3) My lab manager mentioned something about "locking the scale" when it comes to the voltage waveform display. What does he mean by that? He said that when the scale is locked, the display won't bounce around as much..... How do I set the chart options to lock the scale?

4) We are mainly interested in the magnitude of the acceleration at this point and not interested in the direction nor the frequency components. I am reasoning that all that needs to be done to convert voltage to acceleration is to divide the raw voltage collected by the sensitivity of the accelerometer, which is given in (mV / g). Is there more to it than that? Furthermore, I want the Max Magnitude of the acceleration (collected during the entire individual sample run) to be displayed on the front panel. Right now, I made a display for the "instantaneous acceleration" which is just the number I described above (diving by the sensor's sensitivity). Is there a way to obtain the max acceleration magnitude for the entire duration of the sample run? Do I need to send all the instantaneous acceleration values to a 1D array and then from there perform some "Find MAX" function (which I am not sure exists in LV 8.6) and then send that to the front panel indicator? Or is there a much easier way?

Thanks so much for the help! I love these forums!

[jschwartz]

Hi Fascinated,

To answer some of your questions:

1) You can remove noise by using filters - either externally or in software. 

2) You might try using a waveform chart, rather than a waveform graph.

• Types of Graphs and Charts
  

3) You can turn autoscaling on or off by right clicking on the graph then selecting Y Scale >> Autoscale Y.  I imagine that this is what the lab manager was talking about.

4) I would keep it in a voltage waveform and use the Amplitude and Levels Express VI to measure the peak value then multiply this by the sensitivity.  You can find this by searching, i.e., right click on the block diagram and select search