03-01-2006 07:43 AM
03-01-2006 09:42 AM
03-02-2006 08:19 AM
I just wanted to reply to Dennis's post. It depends on which version of DAQmx you are using how the channel spacing works. If you are using 7.4?? or greater (somewhere around there) then the channel spacing is NOT spread out evenly across the entire sample period. Depending on your sample rate it is more like spread evenly across half your sample period. There are other more complex heuristics I can post here if people are really interested.
StuartG
03-02-2006 08:42 AM
Stuart,
Thanks for the update. I haven't had to do a DAQ task since I upgraded to DAQmx 8. I'll keep the changes in mind. I'd be interested in knowing more.
03-03-2006 08:37 AM
I thought you might want more info. So the help for AI convert rate property does somewhat describe the behavior. What the help states is "By default, NI-DAQmx selects the maximum convert rate supported by the device, plus 10 microseconds per channel settling time. Other task settings, such as high channel counts can result in a faster default convert rate." So lets look at an example
Let's assume your board is 1MHz max rate. Therefore the MAX convert rate is 1us. Let's also assume you have 2 channels. This is 22us given the definition above or around 45kHz. So if you set a sample clock rate of 1kHz you are well within this range. We would only use up a small fraction of the total sample time to do the convert period. 22/1000 us or 2%. However as you increase the sample clock rate the convert period starts taking a greater percentage of the time. Finally if you set your sample clock rate to say 50k if we kept the algorithm the same you obviously wouldn't be able to convert all the samples for both channels in this period (50k is 20us and it takes 22us). So, in that case we remove the 10us settling time period so the conversion will happen within the given sample clock period.
I hope this all makes sense
StuartG