Labview Waveform Type into Diadem

[PeterKB]

We have created some TDMS files in Labview using the native NI "Waveform" data type (this is basically a cluster, with the Y axis data, and a Start-time and delta-time structure to document the time axis).

To our surprise, importing this into Diadem seems to ignore the time part of this cluster, and just create our Yaxis structures.

Am I missing something, surely as part of the NI toolset, Diadem should automatically create the required time vector?

I have tried to use the generate time function in Analysis, but it looks like that would require us to manually enter the start time and delta time.

[ML927]

Hey Peter,

Can you share your TDMS file or at least a file that reproduces the problem with dummy data ?

Note that, when importing waveforms in DIAdem it only creates one channel with the waveform type. The time information are stored in the channel properties.

[gsklyr]

Maybe this:

ANALYSIS (tab) -> Chanel Functions -> "Numeric Channels <-> Waveform Channels"

[PeterKB]

Thanks so much for this reply. This gets us one time channel, converts a single waveform to a numeric channel.

We have to use the Relative Time reference, trying to use Absolute won't work.

This may be related to

"The time values of a waveform channel are relative to the absolute start time of a measurement. This means that DIAdem calculates and displays waveform channels with a relative time reference. However, DIAdem does not include the waveform property wf_start_time."

After our conversion, we end up with one time vector, one Waveform that is now a numeric channel, and the remaining waveforms still as waveforms. Despite this we can now properly display the data in View.

However, since we have multiple Groups in our TDMS file, all starting at different times, but can't look at any timing relations, as Diadem insists on treating them as starting at the same time. 

I find the way Diadem handles this unbelievably obtuse, and can't believe that well a structured, self documenting TDMS structure generated in NI tools, can't come be imported cleanly into Diadem. We will have to reevaluate whether Diadem is the tool for us to use, do we just have unrealistic expectations or unusual requirements?

[PeterKB]

Further progress.

It turns out our original problem with the View windows was related to use using View windows that had already been opened by the Example in Diadem (even though we had deleted the data), it looks like some residual property of the window is screwing things up. When we use a fresh window we can get our waveforms to plot in groups correctly.

The remaining issue for us is that all plots start from time zero, so we cannot overlay or plot waveforms that have different start times in a relative manner.

[gsklyr]

There is a post that talk a little about the waveform offset in TDMS and DIAdem:

https://forums.ni.com/t5/DIAdem/Waveform-x-offset/m-p/3787999#M25028

[PeterKB]

Thanks for that. It looks like we could make it work with offset, but it is irritating that we need to workaround the tool, rather than it doing what we want to do out the box.

[ML927]

It is a bit hard to fully help you because

1. We don't know how you save your data in a TDMS in LabVIEW, you haven't shared your VI
2. We don't have any the said data (or a set of dummy date reporducing the problem) to play around with it in DIAdem and find a better solution.

Can you share anything please ?

[PeterKB]

Sorry our corporate policy prevents me from sharing data.

We will continue to play, and see if we can work around the issue.

Thanks for the help.

[ML927]

I did a test, creating a TDMS via LabVIEW 2017 to then open it with DIAdem 2017 and it works properly. My waveform is still a waveform. Which functions do you use in LabVIEW to create your own file ?