time offset on x scale

[aartjan]

I use Labview to save waveforms to a TDMS file that I want to process in Diadem. Unfortunately, Diadem doesn't seem to recognize the timestamp in the waveform to properly scale and offset the time domain waveforms when viewing them. I fixed the issue of offset by writing the property "wf_start_offset" with the first timestamp. Still the x scale will not properly scale. The offset turnes up in a number, not in time.

Another approach is to make a full timestamp channel. What a waste.

Anybody have a suggestion on how to do this smarter?

[Brad_Turpin]

Hi Aart-Jan,

You have indeed found a chink in the LabVIEW-DIAdem Connectivity armor.  The root of the problem is that LabVIEW can handle explicit date-time values with up to something like femtosecond accuracy, since LabVIEW devotes 16 bytes to each date-time value.  DIAdem, however, only dedicates 8 bytes to each date-time value, resulting in an effective date-time resolution of 0.0001 seconds.  When DIAdem implemented waveform channels in DIAdem 10.0, R&D decided to categorically ignore the date-time content of waveform channels.  The "wf_start_offset" property does show up in DIAdem, but it is interpreted as LabVIEW intended-- as a numeric time shift.  This means that your waveform can start at -1.5 seconds for pretriggered data or at -10 degrees for angular data.

The absolute start date-time value of each waveform channel can be read out by a DIAdem VBScript-- but only to 0.0001 second resolution.  This absolute start date-time just doesn't show up visually in the Data Portal, and it is not used to plot curves in VIEW and REPORT.  Your options are therefore to create a separate date-time channel in LabVIEW, which artificially inflates the size of the TDMS file, or to create a date-time channel in DIAdem after loading the TDMS file, which requires running a VBScript.  The good news there is that DIAdem 10.1 can have custom loading routines so that this could be made automatic for your TDMS files.  But either way you will need a separate date-time channel in DIAdem in order to view the data with the date-time information on the X axis of a graph, and in both cases that X axis will have a resolution limitation of 0.0001 seconds.

Ask if you want further information,
Brad Turpin
DIAdem Product Support Engineer
National Instruments

[greentea2]

Hi All,

I am having difficulty with the time scale (x axis) of data from a pressure sensor.  Due to noise, I have snipped a section of the waveform and now need to move the data from a certain time e.g. .4 seconds to start at 0 seconds. Note, I have two channels, one for the time in the following format:

02/30/2011 19:36:51.3521 & one for the pressure.  For the time format, the first is the date followed by ??? so am a bit confused on how I could possibly use offset correction or another function.

Thank you!

[ThiCop]

hello greentea, (BTW nice nickname)

This format stands contains:

- Date

- Time (hours:Minutes:seconds)

Does this answer your question?

[greentea2]

Thank you, Thierry. It's my favorite drink after coffee, of course.

My problem was with x axis offset correlation.  After much searching, I discovered that the time must be extracted from the channel because when imported into Diadem, time and voltage are coupled (more info can be found at http://digital.ni.com/public.nsf/allkb/C74C08AED367563E862576BF005A53E4?OpenDocument).  

Is there a current Diadem user's manual I can refer to? I found one but it's dated few years back.

[ThiCop]

Hello greentea,

One recent Manual you can refer to is for example this one:

https://www.ni.com/docs/en-US/bundle/data-mining-analysis-report-generation-with-diadem/resource/373...

Another useful page is this one:

https://www.ni.com/docs/en-US/bundle/diadem/page/diadem/diadem_main.htm

One other interesting thing might be the DIAdem Basics and DIAdem Advanced courses.

If there are other things you would like to know, then don't hesitate to let us know.

[greentea2]

Thank you! The information you provided has been extremely helpful.