How does Labview measure dt when combining waveforms? Possible mistake in the NI examples

[eli.barber]

In the example project NI 9227 Getting Started.lvproj, the example shows how to read a buffer of single point values from all 4 channels. For the host VI, the dt is calculated from parsing the string enumerator and multiplying by 1000 to get the samples per second. The text "2.00 kS/s (NI 9227)" should parse then to 2000 which gives us a dt of 1/2000 or 0.0005.

However the documentation shows that the enumerator is per channel, not all 4.

This leaves me with 2 questions:

1. I want to know if the true maximum sample rate with all 4 channels active, would it be 2.5 kS/s? And the correct dt would be 0.0020 instead?
   
2. In terms of absolute time, would each the 4 decimated arrays have to be shifted to be properly centered? See:
   1. 0.0005*(0.5)
   2. 0.0005*1 + 0.0005*0.5
      
   3. 0.0005*2 + 0.0005*0.5
   4. 0.0005*3 + 0.0005*0.5

The NI Example project can be found under "Hardware Input and Output \ CompactRIO \ Module Specific IO \ Analog Input \NI 9227 Getting Started.lvproj "

[ZYOng]

The shipping example is correct.

When reading back from the FIFO, the host is always returning data in a multiple of 4.

[eli.barber]

But that is exactly my point!

1. The FPGA reads the 4 channels, and then writes them to the FIFO Block in loop.

2. For the waveform, every dt value for AI0 should be 1/(2000*4) since it has to read the other 3 channels before can it get back to AI0!

[mcduff · Accepted solution]

---

@eli.barber wrote:

But that is exactly my point!

2. For the waveform, every dt value for AI0 should be 1/(2000*4) since it has to read the other 3 channels before can it get back to AI0!

---

The channels are read simultaneously; each channel has its own ADC. Not sure I understand your question.

[eli.barber]

Thank you,

What threw me off was that I also have and RTD card (NI-9217) and an AI card (NI-9209). All three cards use the same language in the documentation of giving the conversion time/sample rate Per channel. And both of these cards are sequential in timing (At least that is what I have tested them to be).

But like you said, the current card simultaneously. This does seem to be the case when running it.

[mcduff]

You can use LabVIEW to access most, if not all, of your device's properties if you are unsure in the future and don't want to hunt for a manual. See the snippet below for your case.

[GerdW]

Hi eli,

---

@eli.barber wrote:

What threw me off was that I also have and RTD card (NI-9217) and an AI card (NI-9209). All three cards use the same language in the documentation of giving the conversion time/sample rate Per channel. And both of these cards are sequential in timing (At least that is what I have tested them to be).

---

• The datasheet for NI9217 clearly lists conversion times for a single channel and for all channels, and so describing a MUXed ADC. The website lists "400S/s"…
• The datasheet for NI9209 gives "conversion time per channel", describing a MUXed ADC. The website mentions "500S/s"…
• The datasheet for NI9227 describes "four simultaneously sampled … channels". The website mentions "50kS/s per channel"…

So the documents provided by NI clearly describe the behaviour of your hardware!