04-28-2014 08:38 AM - edited 04-28-2014 08:49 AM
Hallo,
ich möchte mit einem Windows PC (kein EZB) kontinuierlich Daten erfassen (2-4min). Als Hardware habe ich ein NI cDAQ-9184 Chassis mit einer NI 9208 Messkarte. Leider hat die Messkarte keinen Onboard-Speicher und darum befürchte ich, dass ich Werte verliere.
Ist meine Vermutung so richtig?
Vielen Dank im Voraus!
Hello,
I would like to measure continues data (2 to 4 min) with a Windows PC (no RT). I use the hardware NI cDAQ-9184 chassis with an NI 9208 DAQ card. The measurement card has no onboard memory and therefore I am afraid that I will lose values.
Is that right?
Thanks!
04-28-2014 03:56 PM
Not entirely correct. While yes there is no data stored on the device, the data is first buffered on the device then streamed directly to the PC that the VI is running on. If you do a DAQmx Read, you will get your data. But remember that you will have to read often enough to keep up with the data being generated. Running one of the shipping examples should help.
If you open up LV, and select Help ->Find Examples...
Then in the NI Example Finder, select Hardware Input and Output -> DAQmx -> Analog Input -> Current - Continuous Input.vi
04-29-2014 01:55 AM
How big is the buffer on the NI 9208 or is there documentation about the buffer size available?
It would be interesting for the DAQmx Read frequency.
Thanks so far!
04-29-2014 08:42 AM
The buffer size is specified on the chassis you are using as the module itself has no buffer. If you look at this thread, you'll see that for one AI task the buffer size was seen to be 12 MB. And remember to schedule reads for about half the buffer size, you'll need the extra time/space for networking latency (depending upon your network).
04-29-2014 12:26 PM
Ok, thanks a lot. I'm wondering that I couldn't find anything about buffering (buffer size) in the documentation of the chassis?!?
Another question about the sample rate, the NI 9208 has 500 S/sec for all channels. That means I have signal upddate every 2ms (fast sample mode) if I use only one channel and 32 ms for all 16 channels? Is that right?
04-29-2014 03:41 PM
Yes, the input characteristics on the specifications says that the sampling mode is scanned. This means that for a single channel you will have 2 ms per sample (500 S/s), or for 16 channels the fastest sampling rate you could have would be 32 ms per sample (500 S/s divided by 16 channels).
04-30-2014 02:34 AM
Thanks for your help!
05-06-2014 07:24 AM
Is it also possible to get a time stamp for each measured value? The problem is if I use the buffer I don't know the time!
A 2D array with measured value and time stamp would be great.
05-06-2014 07:49 AM
05-06-2014 08:59 AM - edited 05-06-2014 09:00 AM
Yes, but I also need the tíme stamp for each measured value...