02-02-2009 03:04 PM
I am trying to use cDAQ-9172 and NI 9234 to collect acceleration and force data from accelerometes and impact hammer. The sampling rate I need is 500 Hz, it is not the minimum sampling rate as told in data sheet of cDAQ 9172. But after I saved the data and read it from the file, I found the delta x (step duration) is 0.000605s, where the relative frequency is 1652.89Hz. When I changed it to even smaller rate, what the duration step kept at 0.000605s. What does that mean? The continuous samples and N samples will have the same results.
Anotherthing is that how to get the force analog input? Since I need to input the sensitivity of the force hammer.
Thanks.
02-03-2009 11:37 AM
It sounds like you are using the module's internal timebase of 13.1072 MHz. According to the Module specifications when using the module's internal timebase, the minimum sampling frequency is 1652 S/s - exactly what you are getting. If you are to specify an external timebase the lowest sampling frequency possible is 391S/s. It is therefore possible to get a lower minimum sampling frequency, but you must rout an external timebase to the module to be used in determining the sampling frequency. This is seen in the product manual of the 9234 : http://www.ni.com/pdf/manuals/374238c.pdf
To do this you have two options: use the cDAQ 9172 internal timebase of 80MHz, 20MHz, or 100kHz. or rout an external signal (pulse train with specified frequency-best coming from an internal counter from the cDAQ-9172). Routing a counter signal will allow you to specify any frequency out on the counter and then enable you to get the 500Hz sampe clock you require. If you were to rout the 100kHz timebase from the cdaq chassis to the module, you can only get a sampling frequency of 391S/s becasue the formula is:
fs = fm/(256*n) where
fs = Sampling Frequency
fm = Master Timebase frequency,
n = integer n = 1-31. Note: even if you specify n = 31 the sampling frequency MUST fall withing the module's specified sampling frequency range
To use one of the internal counters as the external timesource you must do this in labview and this is described in the following KB: It is not possible to generate a counter task on these internal counters from MAX.
http://digital.ni.com/public.nsf/allkb/EEB574335BA0B4EB862572060055E9DD
As for your second question, there is no force task for DSA modules. You must use a custom Voltage with excitation Task and create a custom scale define by the sensitivity of your device or use an acceleration task and specify the sensitivity of your device and that will calculate the acceleration directly. Check the following KB:
http://digital.ni.com/public.nsf/allkb/21D7891D9960C5BF862574EA005ADEE6
Hope this helps
01-19-2012 01:01 AM
Hi Charles,
I am using the same hardware (9172 chassis+9234 module) and would like to get the sampling rate of 391S/s for my test. How could I rout 100kHz timebase from the chassis to the module? I tried it using LV Signal Express but the the Reference clock setting tab is locked to the internal type only.
Regards,
Andy N.
01-23-2012 05:04 PM
In Signal Express, it may be easier for you to use the "Subset and Resample" which can be found from Add Step>>Processing>>Analog Signals>>Subset and Resample. This will take the higher sampled data and generate a new set of samples at any frequency you need. This step can be added to the program you have now and will not require you to modify the timebase of the NI-9234.
Kyle K.
01-24-2012 04:31 AM
Hi Kyle,
Thanks for your reply but using resample/decimation means applying interpolation techniques and/or filters which would change a bit the content of signals (e.g. amplitude) while I need pure ones in my work. Also, we are only interested in low-frequency range and using high-sensitivity sensors, so 391S/s is too enough for us. Please advise me how to make use of the timebase of NI-9172.
Regards,
Andy N.
01-24-2012 04:03 PM
I would strongly recommend oversampling at integer values of your desired frequency and decimating the data (only using the real data that was read at the intervals you want). I recommend this because in order to import the timebase, you will require a DIO module (such as a NI-9401) and it will have to be programmed in LabVIEW, not Signal Express. It will utilize advanced programming methods to generate a pulse train (which is the master timebase we want) and route it to the DSA module. The decimation will still be true measured data which means it will be effectively the same as sampling at the lower rate.
Kyle K.
01-24-2012 05:20 PM
Kyle,
Thanks for your swift clarification. If it is impossible to rout the chassis timebase to the module using Signalexpress and DAQmx, I would have to go with the solution of decimating data.
Regards,
Andy