Trigger 6259 & 6025

[super-neuron]

I have&nbsp;two&nbsp;NIDAQ cards(6259m &amp;&nbsp;6025e).&nbsp;&nbsp; I&nbsp;would like&nbsp;to&nbsp;simultaneously trigger three analog output tasks across these&nbsp;devices at different sampling rates with a reasonable level of synchrony(&lt;1ms).&nbsp;&nbsp; I am new to multiple device synchronization and RTSI(installed &amp; configured) and have had difficultly finding information on this topic.&nbsp;&nbsp; I located&nbsp;the article <a href="http://zone.ni.com/devzone/cda/tut/p/id/3615" target="_blank">M Series Synchronization with LabVIEW and NI-DAQmx</a>&nbsp;which has a section titled "Special case: E Series and M Series Synchronization," however this example only allows for a single sampling rate.&nbsp;&nbsp; Can anyone help me find more information on this topic or advise on the accuracy&nbsp;of the attached VI?&nbsp;&nbsp; Thank you.
&nbsp;
Sincerely,
Zach


example_1.vi:
http://forums.ni.com/attachments/ni/250/41889/1/example_1.vi

[Corby_B]

Hello super! Thanks for your posts and questions.So are right in that you can only use one AO sample clock hardware timing engine per board. So the example you are showing that has two different AO tasks can't acutally be done. So what I would do is put both your AO outputs in the same task of Dev1 and then use the higher sampling rate of the two sample rates you need. Also since you have the RTSI configured you can set it up in the picture I have attached below. If you do need the different sample rates you can always stop your first task and then configure and start a different task. The driver takes care of all the signal routing for you so I am using the start task on one board to start the other. Let me know if this helps and please post back if you have any other questions.Cheers!Corby_Bhttp://www.ni.com/support


example_1.JPG:
Loading Image...

[super-neuron]

Corby_B,
Thank you for your response.&nbsp;&nbsp; Bundling two AO channels(dev1/ao0:1) into a single tasks is a novel approach I had not considered.&nbsp;&nbsp;&nbsp;To clarify, I&nbsp;need to buffer two completely different signals on each of these channels:
dev1/ao0 = N samples of uniform white noise
dev1/ao1 = single&nbsp;5v pulse
Is it possible to&nbsp;'DAQmx Write' two different signals on a single task?
Sincerely,
Zach

[Corby_B]

Hello Super! Thanks for your post back.I am glad to hear that things are going well for you. That is a great question about the 3 tick delay on the AI clock and convert clock. I see in your program that you are doing only AO tasks. Each AO task has its own Digital to Analog Converter (DAC) so there is no Convert clock running in the background. You only see this delay when you are running AI tasks because there is a sample clock and a convert clock and because the AI inputs are multiplexed together with only one Analog to Digital Converter (ADC). Your program will be synchronized. Let us know if this information is helpful to you and I wish you the best on your applications! Have a great day!Cheers!Corby_Bhttp://www.ni.com/support

[super-neuron]

Pardon&nbsp;my&nbsp;last question... I figured it out.&nbsp;&nbsp; I used the build array component to pass an array of waveforms into the DAQmx Write component.&nbsp;&nbsp; Works great and cleaned up my block diagram!
Based on suggestions here and a tutorial I found online called&nbsp;<a href="http://zone.ni.com/devzone/cda/tut/p/id/3615" target="_blank">M Series Synchronization with LabVIEW and NI-DAQmx</a>, I created the attached&nbsp;example_2.vi.&nbsp;&nbsp; I'll be&nbsp;putting on the oscilloscope shortly!
I did however what to ask one more quesiton.&nbsp;&nbsp;&nbsp;In the above tutorial under the section entitled "Special case: E Series and M Series Synchronization," it makes special mention of the fixed hardware delay in the E series default sample clock.&nbsp;&nbsp; In the example given, a DAQmx Timing property node is used to delay task start by 3 ticks of the&nbsp;sample clock to compensate.&nbsp;&nbsp; The example is illustrating AI, and when recreating this&nbsp;for AO, I found I could not implement that&nbsp;delay in the same manner.&nbsp;&nbsp; Do I need to worry about this delay when performing AO?&nbsp;&nbsp; Can I expect the attached&nbsp;example to produce precise synchronization down to a single clock tick?
Sincerely,
super-neuron


example_2.vi:
http://forums.ni.com/attachments/ni/250/41954/1/example_2.vi

[super-neuron]

Thank you Corby_B,
&nbsp;
That is exactly what I needed to know.&nbsp;&nbsp; I greatlly appreciate all your help!
&nbsp;
Sincerely,
super-neuron