2012 Waveform Chart Bug - Workaround?

[mjaspan]

I'm posting this message in search of assistance in finding a workaround.

There appears to be a bug in how Waveform Charts display data under LV 2012.  See the attached VI as an example.

Under LV 2012 the Waveform Charts do not appear to scroll the plot correctly when the plotted data reaches the side of the Chart.  The manner of the problem depends subtly on the timing, and is illustrated by the attached VI.  I've tested the issue under LV 2010 and the bug did not exist then.  I skipped LV 2011, so I don't know whether this bug was introduced with LV 2012 or with LV 2011.  To demonstrate the bug, run the attached VI in LV 2012.  To see correct behavior, save the VI as LV 2010 and run it in LV 2010.

I've spoken with NI tech support and a CAR was filed.  (CAR#381638).   I am now searching for a work-around.  

I am managing a medium-size software application which I transitioned to LV2012 a couple months ago.  

I would prefer to avoid switching back from LV 2012 an older LV version as the work-around, but fear that may be the best solution.

The most obvious workaround, using Waveform Graphs instead of Waveform Charts, is problematic for a variety of reasons, not the least of which is the added code complexity.   My application makes frequent use of the Stacked Plots feature of Waveform Charts.  This would seem to preclude the obvious work-around of using Waveform Graphs as a substitute for Waveform Charts, since Waveform Graphs do not support Stacked Plots.

  I'm using LV 2012 and Win7.

  Perhaps someone with a better understanding of this bug, and/or understanding of the internal implementation code of Waveform Charts can describe better the root cause of the bug and aid in searching for a work around.  The problem is subtle and depends on the execution speed of other parts of the VI.  For example perhaps I need to send the data more frequently, or less frequently, or set a flag, or turn a feature off.... (?)

  Any suggestions for a workaround are appreciated.

Thanks,
  Martin

[The_Hunt]

Hello Martin, 

Thank you for bringing this bug to the attention of our support engineers. Currently I have been unable to find a suitable work around other than the two that you have already mentioned. The only other thought for a workaround would be to use the waveform graphs and set them up in your front panel to appear like stacked plots. You can remove the x-axis from the top graphs and remove other front panel items to make it appear more like one graph. This workaround will require additional coding with the use of property nodes in order to sync the axes. 

I understand that in no way is this workaround ideal, since it will require more effort in order to simulate the stacked features that exist for waveform charts. I will continue to look into this and see if I can find a better work around for you Martin. 

[mjaspan]

Hello Huntington,

  thanks for looking into it.  I have made a few observations about possible workarounds.

I've noticed that the bug sometimes has a bigger scalling error and sometimes smaller.   In fact some of my code does not demonstrate the issue at all.   I'm still unclear on the interactions, but it seems to be mostly determined by loop time or update rate on how often data is passed to the chart.  

Also, some actions cause the chart to re-update itself and correct its display.  For example, each time the chart is moved on the front panel will fix it, either in Edit Mode by using the mouse or during Runtime using Property Nodes.  Unfortunately this solution requires that the chart wiggle and vibrate on the FP, but it gives me hope that there's a less distracting solution.

Alternate workaround -->  I've also noticed that this bug appears to only affect the Strip Chart Update Mode.  The Scope Chart and Sweep Chart Update Modes appear to work correctly, which while not ideal, is at least a work around which doesn't require extra coding.

Again, I am hoping that others affected by this bug will reply here with their observations which may help tease out a pattern in the interactions which can be used to suppress this bug.  At least until a fix is relased, presumably in a future version of LV.

thanks,

  Martin

[mikeporter]

Here's another take on it. It isn't pretty but depending upon your chart update rate you might be able to live with it, plus it will be easy to take out when the bug is fixed. The only other thing is that is you don't want to have to buffer all the values, you will need to also update the graph XAxis start value with each iteration after the buffer is "full".

Mike...

[dan_u]

I just encountered the same bug in LV2012SP1 (with Patch f3). It's really a serious bug since the time axis does not match the data.

It was reported more than a year ago and there's a CAR.

Has there been any action since? Any patch, workaround (other than using a graph)?

[Jeff-P]

Hi dan_u,

There has not been a patch to address CAR 381638 in 2012 SP1. I can confirm that we were able to find the bug and it is fixed in the upcoming release of LabVIEW 2013. 

Regards,

Jeff Peacock 

Product Support Engineer | LabVIEW R&D | National Instruments 

[dan_u]

Thanks Jeff for the info. I'm not going to start another discussion about this bugfix policy (you can only have bugfixes along with new bugs in a new release).