Which software, c++ or labview..., for high time-resolution( µs)?

[Alex666]

Hi,

i need to measure data with a time resolution of less than a few microseconds. Is this possible with labview? Or will i have to use C++?

The interface cards, i could use are: CB-68LP or BNC-2120 both from national instruments

Thanks

  Alex

[Jochen]

Alex,

As the data acquisition is not timed by the software but by the hardware that you are using it doesn't matter which programming language you choose for your task. By the way, you have just provided information about the external connector blocks but not about the data acquisition device that you will use

LabVIEW is capable of analyzing continuous data streams with several MB/second which is comparable to C++ so I don't think that this will be the main criteria to decide which programming language to choose.
You should choose whichever approach is more appealing to you (graphical or text oriented). In my opinion the graphical approach is more intuitive and faster and it's easier to maintain code that was written some time ago but I don't want to start a debate on principles with this statement.

If you prefer the text based approach please have a look at NI's Measuerement Studio which provides great features for user interface design, signal analyzing and data acquisition.

Best regards,

Jochen Klier
National Instruments Germany

[Alex666]

"timed by hardware" means the speed of the computer i´m using? Or are there some differences between the different boards? The board i´m using is the  PCI -MIO-16E-4 and the PC is a AMD/Duron with 900MHz. Is this enough for µs-resolution? Or ,if not, which hardware-setup would be usefull for this task?

[Jochen]

No, hardware timing means that the data acquisition device works with a specific sample rate that is independant from the PC that is used. The sample clock is derived from an onboard oscillator on the PCI-MIO-16E4. The maximum sampling rate of this device is 250 kS/s so the maximum time resolution that can be reached is 4 µs.

Jochen

[Alex666]

i´m not sure if i expressed myself really clear( probably not). i don´t have to measure one data-point with a high time resolution. i have to measure a very long sequence of data-points. so the time distance between every datapoint should be 1 µs or less. is this really independent of the PC-speed

sorry for the confussion

Alex

[Joe_H]

Yes this is independent of the PC speed. and with the card that you have the best timing you are going to get is 4uS as stated previously.

[Alex666]

Thanks for your hints, i´m one step further.

Now i face the problem, that i want to read the same analog input data with two different time-resolutions in one program. The one data acquisition must be slowed down by the wait until next millisecond multiple, the other should be at 4µs.

when i try it with two seperate while-loops an error occurs: no transfer is in progress for the specified resource #10608

i also tried it running two seperate programs and the same error occured

can you help me?

thank you once more

 Alex

[Jochen]

The PCI-MIO-16E-4 has only one AD converter. All channels are multiplexed to this ADC and they are transferred through a single DMA channel to the host memory. Thus you can run only one acquisition at a time. That means you can run a hardware timed acquisition with up to 16 channels at the same time but alle these channels need to be acquired with the same sample rate. You can't run a second analog input task in parallel.

So for your application you should read all required channels in the same task. If you don't need all data from a channel then simply thin out the data after acquiring them. Please note that the sample rate per channel is the maximum sample rate of the board (= 250 kS) divided by the number of channels used.

Best regards,

Jochen Klier
National Instruments Germany