Acquiring many signals with multi devices (NI 9229 and NI 9205) trough a cDAQ chassis 9188

[phidu]

I use a cDAQ chassis NI 9188 equiped with 6x NI 9229 (6x 4 AI chnls simultaneous sampled with aliasing filter) and one NI 9205 (32/16 AI chnls single/differential multiplexecd chnls). I have created a simple task wit the DAQ Assistent to acquire 12 chnls with the NI 9229 modues and 4 with the NI 9205 module. A generator is used to produce a sine wave with a frequency of 10 Hz and an amplitude of 500 mVpp as input to all channels. The same generator produce also a square wave (synchron to the sine) which is used as trigger input (PFI 0 - BNC connector on the chassis). The acquisiton is set to take finite samples by a data rate of 2 kHz and a datalength of 4000 samples (it means a acquisition time of 2 secondes) controlled by a digital reference triggered with the cADQn/PFI0 as trigger source.

My problem is the time difference between the signals acquired with the NI 9229 and those acquired with the NI 9205 (show the attachment). These time difference is not constant and change when the frequency of the input signal varies. I suppose that these difference is due to the aliasing filters used in the NI 9229 and I am searching how to resolve the situation. Any help will be appreciated...

Philippe

[Stephan_D]

Hi Philippe,

I don't know if this relateted to your issue, but please keep in mind that the NI-9205 only used one ADC converter. All the inputs are routed through a MUX to this ADC (http://www.ni.com/pdf/manuals/374188d.pdf, page 9). So you will receive a delay on every sampled channel which is not avoidable. It is not possible to elimate this problem properly in software.

The NI-9229 is a module with a ADC for ervery channel. This modules should work like expected without any delay between the channels.

Regards, Stephan

[phidu]

Hello Stephan,

Thanks for the feedback. You're right and I known that we acquire the data over a mux but I am really surprised about such time difference, according to the fact that the acquisition is started at the same time for all channels on all devices with a scanrate of 2 kHz. The specifications for this module shows a settling time of 4 or 8 µsec, so in my opinion, this should not generates such difference between the channels...

Important to known is that the time difference varies if I change the input signal frequency to be, sometimes, equal to zero.

I agree that with a device like the NI 9205 we will have a time interval between the channels but an interval in µs is for our application acceptabel, interval in ms not more.

I'm maybe wrong but I would find great to read other news about this subject.

Regards, Philippe  

[phidu · Accepted solution]

By searching and reading the NI KnowledgeBase, I resolve the problem by performing a dummy read on the NI 9229 channels before reading those from the NI 9205 device.

According to the specifications of the NI 9229, there is a filter delay (digital filter) of more or less 40 samples which are present only in case of a digital triggering.

Reason and solution of this problem are explained here: http://digital.ni.com/public.nsf/allkb/F989B25FF6CA55C386256CD20056E27D.

I have apply the solution presented in this document and it's work great.

Now, this problem is resolved.

Regards an all, Philippe