Waveform Triggering - APFI Versus AI

[StephenB]

No existing DAQ PXI or PXIe products support multiple different AI sample clock speeds. All the ADCs share a single wire on the PCB for sample clock. So that's why you define waveform or single point mode on the entire AI section.

However, there may or may not be future DAQ products that handle this differently 🙂 NIVS would of course need an update to take advantage of that

[StephenB]

Steve,

If you can... Can you describe why you are doing high speed data acq in parallel with HIL? We typically see single point and waveform combined in Test cells, not traditional HIL. Is this a test cell and an HIL system combined? I just want to understand the market better.

Thanks!

[Pie566942.0]

Thanks for reiterating Steve, I do understand.  I was confused because I wasn't aware of the drop-down to configure the AI channels as SP or waveform.  I only looked at the individual channels.  When I added the AI channel to set the input configuration and range (but was not aware to configure the AI group to waveform), and also had the channel defined in my waveform task, I could still map the waveform channel.  I couldn't reconsile the two.  I'll PM you on your other question so I don't get fired.  Thanks guys!

-Steve K

[thumble]

Hello Steve,

If I may say so, for some years now, most of the test benches we had to realize for Aerospace and Automotive are mixing the HIL, calculation and data acquisition functions in the same hardware.

Best regards

[StephenB]

Thank you for the info. Is this because you are doing more of a test cell style application with physical phenomenon being measured? Or is it still electronics HIL that needs waveform acquisition?

[thumble]

Hello,

It's typically because the same input signal needs to be recorded at HF, as it is also processed in RMS calculation to be used in real time by the model(s), monitoring functions, MMI display etc... More generally speaking, the same input is recorded at HF and used in real time processing.

[StephenB]

Interesting!

Do you write your own waveform processing custom devices for this?

[thumble]

Yes we did because we had to record an overall flow of more than 10 000 000 samples/second per node, coming from different DAQ boards, including any required VeriStand channel and PTP synchronized with other nodes.

The philosophy for this board is to acquire at HF the AI channels of a series X or M board (in raw values for data size issues), push this flow into a shared ring-buffer (allowing other subscriber to perform HF calculations i.e. RMS) and in a recording buffer (allowing a recording task to work). This board deals with VeriStand channels: the DI, DO, AO are treated classically (they are not oversampled), but for each AI there are 3 channels that delivers the maximum, minimum and latest value for the current VeriStand loop.

[StephenB]

Ah ok. I think I know what project you are working on then 🙂

We almost had time to put waveform max/min/latest value into NIVS 2012 but we ran out of time. Really wish we got to it 😕 would have been really nice