08-23-2011 05:34 PM
Since I've had a chance to fully experience Signal Express 2011 on our "lab PC" with real-world test applications I've discovered a few reproducible issues.
I'll post these and all future issues here; and I encourage anyone else who experiences unexpected behaviors, bugs, hangs etc should post in this thread as well. During NI Week had an opportunity to speak one-on-one with a few members of the Development Team. A lot of great feedback was exchanged but one thing that was expressed was the value of having consolidated threads like this one and the "Improvement Ideas" thread. It really helps them get an understanding of user needs and facilitate priorities.
Don't let your frustrations go unheard. The Dev Team really does read these comments and they take our suggestions very seriously.
Next Available File Name, Per Iteration:
This seems like a really useless function but with this option selected in the Save to ASCII step SigEx creates a single txt file per data point.
Example: I recently took 20 seconds of data at 1KHz with this option (accidentally) selected. At the end of my acquisition I was left with 20,000 .txt files in my project folder, each containing a single sample point worth of data. Deleting 20,000 files locked my PC up causing me to hard restart.
Scalar XY Plots break when opening an existing .seproj file.
When launching SigEx, then opening an existing .seproj file which contains xy plotted graphs, the graphs will not display any "live" data. The fix is to turn the xy graph into a different display type; i.e.. right-click, View As... gauge. Then right-click, View As... change back to scalar xy graph. Graph will now show live data
Abnormal display when viewing raw signals as scope or sweep chart
When viewing raw signals (from DAQmx Acquire step) in a waveform chart and choosing Scope or Sweep Chart as the update mode, only a small portion of the full graph is used. Its as if there's an invisible cutoff point in the graph. The data will sweep within this small portion, but it never scrolls across the whole graph.
This is only detectable in Scope or Sweep Chart mode. Strip Chart is unaffected.
Cursors do not follow zoomed-in signal in Scope Chart mode
This has been addressed in an earlier post and it's determined to be a definite bug. Just figured I'd consolidate it in this list.
I'll post more as issues become discovered.
08-25-2011 07:04 AM
Thank you for this. I'll try to reproduce all of these issues and make sure to bring them to the developers' attention. It's especially helpful to us that you listed the steps necessary to reproduce and workaround(s) you've already discovered.
09-02-2011 08:23 AM
Hi OKors,
I wanted to give you an update on these issues. I was able to replicate the Per Iteration, Scalar XY Plots, and Scope/Sweep behaviors, and checked on the status of the cursor in Scope Chart mode.
Next Available File Name, Per Iteration
CAR #298421
Although this behavior isn't necessarily a bug, I do see how creation of an extraordinarily large number of files could be a problem if there is no warning, especially if done in error and deletion of so many files crashes the computer. A CAR has been filed on changing the help file to specify "per iteration" in the Save to ASCII/LVM entry to help clear this up.
Scalar XY Plots break when opening an existing .seproj file
CAR #313498
Definitely not intended behavior. I filed a CAR on this issue with R&D and also provided them with your workaround of changing the graph type to something else and back again.
Cursors do not follow zoomed-in signal in Scope Chart mode
CAR #299257
This issue is currently being evaluated by R&D, but unfortunately I do not have a timeline on when this will be fixed. I can tell you that they have been looking into it and all that it entails (see below).
Abnormal display when viewing raw signals as scope or sweep chart
CAR #299257
R&D believes that this issue with the Waveform Graph in scope/sweep modes is related to the underlying cause of the cursor behavior in the Scope Chart that you brought to our attention previously, and hope that a fix to one will fix the other. Thus, this now shares the same CAR #.
Thank you again for taking the time to put all of this together, and please let us know if you come across any other issues!
11-04-2011 11:27 AM
SignalExpress 2011 Cache Issues
I'm creating a User step in LabVIEW. When I build it, I add it to SignalExpress as "Run LabVIEW VI", Select this VI from built LLB and chose to run once. It works, but I made a mistake in the VI and I go back and fix it in the project, rebuild, then I open SignalExpress and add it again, run it and it still runs the first version with the mistake. I can see in the Step Setup that LLB was updated, some of the VIs in LLB have changed, but when I run the step, it always runs the first version that doesn't exist anymore. Relaunching SignalExpress and creating a new project has no effect. I guess, that it caches the code and subsequent reopening of the step with the same name doesn't have any effect, it still executes the code from the step when it was added for the first time. To confirm this, I deleted the files in C:\Documents and Settings\All Users\Application Data\National Instruments\Assistants and added the step again, this time the current version was executed.
I've already reported another issue with cache, when I installed SignalExpress Tektronix Edition 2.5.1 beside SignalExpress 2011 and Tektronix steps didn't appear in SignalExpress 2011, this article helped to fix it:
http://digital.ni.com/public.nsf/allkb/DFACEDFC86979CD08625777A00662503
Alexey
11-08-2011 12:36 PM
Alexey,
Thank you for alerting us to this I will go ahead and take the appropriate steps to try and reproduce what is occuring and will forward the issue on if needed.
11-11-2011 01:25 PM
Alexey,
I'm not seeing the issue that you are describing. I am able to build a Source Distribution in LabVIEW that generates a sine wave and view it in Signal Express. I created a project and opened the vi to display in the plot area. I save the project and close Signal Express because I can't rebuild the Source Distribution while the project is open. Then I go back into LabVIEW change the wave to a triangle wave. Rebuild the Source Distribution. And reopen the Signal Express project and I see the change.
Attached is a screen shot of the vi I used. If this issue is still happening for your I could take a look at your code and try it on my computer.
07-31-2012 05:05 PM
We've recently acquired a new laptop and Compact Daq chassis so I'm starting to use parts of Signal Express I've never used before. Here's what's going on:
Using a cDAQ 9172 and NI-9237 module to scale a full bridge load cell.
When entering values in the Create Bridge Scale settings, the settings do not always apply after clicking OK to close the window. Rather I've had to go back into the Configure Scale... window and re-apply my settings. The values usually save after this second attempt but for the unknowing user they would likely go about their buisness with a transcuder scaled to whatever the default values were.
Steps taken:
This is on a Dell Core i7 Latitude running Win7 and Signal Express 2011.
I also notice this happens with the Pressure (Bridge) settings as well.
08-22-2012 03:39 PM
This may not be a bug but I figured I'd post it anyway because it's unusual behaviour and I scratched my head for a brief moment over it.
It appears you cannot add a waveform signal to a graph displaying a scalar dbl signal and get expected results.
I'm acquiring a load signal and passing it through an Amplitudes and Levels step.
I've selected the DC (dbl) channel and dropped it into a graph for view as a time-history Scope Chart.
Now I want to view my min/max's so I add a Statistics step to the setup and leave the default waveform channel as the Input channel.
Drag the min/max values into the graph and take data.
The min/max values "follow" the DC load signal. They do not hold at the min or max values as expected.
This occurs when the waveform data is selected as an input channel to the Statistics Step.
Scratched my head for a minute then changed the input signal to the Amplitudes and Level's DC signal and the Min/Max graphs hold at the appropriate levels.
Pictures worth a thousand words so see the attached screenshots for a better visual.
The first one is graphing dc load and waveform min/max
The second is graphing dc load and the min/max from the dc step.
08-22-2012 03:49 PM - edited 08-22-2012 03:52 PM
Come to think of it...
It appears you cannot view waveform min/max data AT ALL; let alone on the same graph as the originating waveform signal itself.
If I drag the same Load signal (raw waveform data) into the graph and view it as a waveform chart (omitting the A&L step entirely), then add a Statistics step, it automatically turns the min/max into "dbl" signals. If I try to drag the new min/max dbl signals onto the waveform chart created above it splits the graphs into seperate charts for me. It will not allow me to view the waveform data, and the min/max together in the same chart.
11-08-2012 03:33 PM
I've found another highly reproducible bug which is a total show stopper for the unsuspecting victim.
Turns out if you create a SigEx project complete with data view charts, xy graphs and numerical displays. Then go back and rearrange the channel order within the DAQmx Acquire Channel settings. The rearranging of channel orders "breaks" the association between the channel name and the data signal.
It appears that all its doing is rearranging the name but leaving the signal (or visa versa... hard to tell what's actually going on).
What's most concerning is once you've rearranged a few channels then go back and view the graphs or numerical displays in the data view, the signals which show up in the graph or numerical display is from another channel.
The only way to reassociate the channel name to the signal is to close the project and open it up again.
Hopefully these screenshots will aid in a more visual understanding of what's going on but this is serious guys!
I had a bit of an embarrassing moment when my Engineer who made a trip up from Division wanted his signal order to be mixed around and when I ran the test my data view graphs were all hosed up.