Stopping a (Begin/End)ReadMultiSample()

[densmore]

I have setup a task with AI channels and timing parameters. I've have the timing to read a finite number of samples. I'm using the endreadmultisample(ar) function to read the data back and then check if the task.IsDone to determine if I should call BeginReadMultiSample. See the attached file for the complete code. My question is -- Is there a more effiecient way of stopping retievial of samples after the task has finished?

[nd]

Hi densmore,
Do you want to read a continous amount of data or finite ?
I have to admit that i'm a little confused about why you are reading 2 points at a time.
(though this code may just be a snippet of what you want to do...)

imo a more efficient way to do it would be to specify 10 points in the ConfigureTiming call, and then specify 10 points in the BeginRead (not 2). Whenever your 10 points are read, the callback is called...
Since the BeginRead is asynchronous, (and the actual read occurs in a different thread) really the amount of data that u are reading does not matter, in that the UI WILL remain responsive thu out the operation..(if that is a concern).

This way you do not need the myTask.IsDone block and you do not need to call any additional BeginRe
ads (you'll have just one instead of 5).

The AcqVoltageSamples_IntClkAnalogStart example,that ships with MeasurementStudio (MeasurementStudio/DotNEt/Examples/DAQmx/AnalogIn/MeasureVoltage )would be a good reference, without the triggerring of course..

If you want to read a specified amount of points continuously (say 10 points continuously) then you don't need to set the number of samples in the ConfigureTiming. You can set the ConfigureTiming to be Continouos (instead of finite) and then specify the numberOfSamples to read as an argument to the BeginRead..
Hope this helps, if it doesn't 🙂 or if you have any more questions then please feel free to email us again..

Best regards!

[densmore]

Thanks for your response. What I want to do is take ~4000 measurements over a time period of between 30-300 seconds. While taking the measurements I want to be able to graph the results and do other various analysis of the signal. So basicly I want to view the results in "real time".

[TonyH]

In NI-DAQmx, the data acquisition (measuring the voltages) and the data retrieval (reading the values) operate independently. Task.IsDone determines whether all of the data have been acquired, but when it reports true, it is possible that you have not retrieved all of the data even though the DAQ card has acquired them. Because of this, you should not use Task.IsDone to determine whether to continue reading.

In a finite acquisition, you have set a particular number of data points to acquire. The best way to stop is simply to stop reading once you have retrieved all of the points you set the task to acquire. You can use Task.Stream.CurrentReadPosition to determine how many points you have read, or you can simply keep the count yourself.

What is happening
is that once the task starts, either by calling Task.Start or by doing your first read, the NI-DAQmx driver starts measuring and putting data into a buffer at the rate you request. Whenever you read 2 points, the read asyncCallback will be called and EndRead will return the first two points that you have not yet read. If there are not two points you haven't read, it will wait until there are before calling the asyncCallback for the read.

Incidentally, if you attempt to read again after the task is done, it will simply restart the task.

Let me know if you have any more questions.

Tony H.
Measurement Studio

[nd]

From your explanation, what you are doing makes sense..there is a trade-off of course, between efficieny and getting the results in "real-time"..

One way i can think of, that may be a little more efficient, is to still avoid the check for myTask.isDone (the call to the driver).
Instead you can make it a continuous operation, and use a variable to calculate how many points have been read (you are specifying the number of points to read in the Read-use that). Obviously , the bigger chunks you read, the more efficient your operation will be.
When you have read the requisite number of points, you can call a myTask.Stop.
This is of course based on the assumption that quering a variable is quicker than querying the driver, which is a fairly good one
🙂

Hope this helps, please feel free to post again if you have any further queries...

[densmore]

Thanks for all your responses, it helps a lot. I have a second question. I setup a task with AIChannels, timing, and a trigger. I start the task, and the board receives the first trigger. It aquire's and process's all the data. Then the board receives a second trigger. Does the task restart? Does the CurretnReadPosition reset to zero? Basically I want to preform the same task multiple times without having to restart my DAQ program.

Also are there any books out for programming in MS using C# and .net framework?

[TonyH]

In the case you describe, the task will not start acquiring data again when the trigger occurs again unless you start the task again once you have retrieved all the data from the first run. You can restart the same task as many times as you want without restarting your program.

Starting the task is something that must be done in your application by calling Task.Start or performing a read operation. Once the task is started, it begins watching for the start trigger, if any, and starts acquisition when the trigger occurs. Once all of the samples have been acquired, the task stops and the data remains available for you to retrieve until the task is started again.

Also, if you start the task using Task.Start, it is best to always
use Task.Start to start the task and call Task.Stop when you are done. If you start the task by reading from it, you should always start the Task in that way.

As far as books go, there is nothing specific to Measurement Studio so far, but there are an abundance of good books and websites for .NET programming. For books, look at Jeffrey Richter's "Applied Microsoft .NET Framework Programming" and for websites, MSDN and GotDotNet are good places to start.

Tony H.
Measurement Studio

[densmore]

I'm kinda concerned about missing the next trigger, because I'm still reading the data from the buffer. When a task is restarted the currentreadposition is set to zero? Is the data buffer emptied when a task is restarted?

[TonyH]

You are correct. When a task is restarted the current read position is set back to zero and the data buffer is emptied to prepare to receive new information. However, there is another way to do what you want.

You can use a counter and a continuous analog input task to guarantee you don't miss any triggers. I've attached code that demonstrates how to do this on a MIO 16XE 50. (If you have a different card, you may need to adjust some terminal names for your hardware, but the overall steps are the same).

I was also able to reproduce the problem you were having with the graph not update. It was due to data being kept in the buffer on your DAQ card and not being transferred to the computer. To force the on-board data
to be transferred to the computer more regularly, you can either force the card to flush the buffer whenever it has the opportunity like this

aiTask.AIChannels.All.DataTransferRequestCondition = AIDataTransferRequestCondition.OnBoardMemoryNotEmpty;

or you can reduce the size of your buffer:

aiTask.Stream.ConfigureInputBuffer(2*count);

where count is the number of points you read each time.

Let me know if you have any problems with this.

Tony H.
Measurement Studio

[densmore]

I added the

AIDataTransferRequestCondition.OnBoardMemoryNotEmpty;

property and that solved the problem. Was the card not transfering the data form the fifo to the software buffer fast enough? I was able to use the debuger and see data in the array that I wanted to plot, but it wasn't showing up on the waveform. Thanks for the help. Really GOOD tech support!!!!!!!!!

JMD