ActiView exes hang

Post Reply
msergi
Posts: 5
Joined: Tue Sep 16, 2014 6:49 pm

ActiView exes hang

Post by msergi »

I am having issues with ActiView crashing on my target machine. Initially I was building an exe on my development computer (slightly modified version of ActiView 7.05) and it runs fine there. When I run the executable on the target computer it will either crash (reports not responding then exits after a bit) or it the NI error handler will report an access violation then exit. The same behaviors occur if I build an exe using the vi downloaded from the BioSemi website. This was using LabView 2013 SP1 and the corresponding run-time engine. The events typically happen during state changes, such as starting or stopping communication with the hardware, but sometimes while running or logging.

Today I installed the 8.6.1 run-time engine and ran the exe downloaded from the website. I click start and can view/log data with no problem. If I stop then click start again the UI immediately hangs. Windows will report the application is not responding and eventually close it. This happens 100% of the time on my target machine. I tried uninstalling all NI software on the machine and re-installing just the 8.6.1 run-time engine, but this did not help.

The target machine is a Surface Pro 3, the i5 version, with 64bit Windows 8.1 Pro. For reference the development machine is a Surface Pro 2. I will try to test with another Windows machine to see if I can glean any information from the success/failure...

If anyone has any insight or suggestions I would appreciate them.

Thank you,
Michael

Coen
Site Admin
Posts: 1124
Joined: Fri Mar 26, 2004 7:00 pm
Location: Amsterdam, Netherlands
Contact:

Re: ActiView exes hang

Post by Coen »

We have heart a few incidents were there seems to be a problem when compiling the original 8.6.1 VI into LabVIEW 2013. The problem shows itself as intermittent Synchronization error (error 5002) popups in ActiView. Your problem seems to be different. To be sure, please confirm that you load the original (and modified) ActiView VI into LabVIEW 2013, and that the program then runs without errors.

Please confirm that the LabVIEW 2013 development computer is also running Windows 8.1

Does the Syntest program (in driver package) run without errors on the target machine ?

Does the problem also occur with ActiView Light (http://www.biosemi.com/download/ActiView604-Light.zip) ?

Does the problem also occur with the LabVIEW BDF reader (http://www.biosemi.com/download/ToolsFo ... er-808.zip) ? If no, then it is most probably related to the USB driver. If yes, then it is a pure LabVIEW problem.

Best regards, Coen (BioSemi)

msergi
Posts: 5
Joined: Tue Sep 16, 2014 6:49 pm

Re: ActiView exes hang

Post by msergi »

Coen,

I can confirm that both computers are running Windows 8.1. I have been able to run both the original and modified VI's in LabView, as well as compile the modified one into an exe and run it on the development computer without issue. I have not tried running the original vi, compiled with LV2013, on the development computer however the modified one works so I expect the original to work as well.

I will try using synctest and ActiView Light later today and report the results.

msergi
Posts: 5
Joined: Tue Sep 16, 2014 6:49 pm

Re: ActiView exes hang

Post by msergi »

I did not have a chance to compile and run synctest, but I did compile an exe for ActiView Light and ran it on the target computer that I have been having issues with. ActiView Light had no problems... I tried starting and stopping communication and the application many times and it did not hang like ActiView does. I re-tested ActiView again (not Light) and it still would hang as before. One new behavior is that twice I observed ActiView report a 5000 error (could not communicate, check power and the cable). I don't know if that helps give any insight given that normally it just hangs or reports an access violation.

I have more testing planned this morning, I will post any new findings.

Thank you again for any assistance,
Michael

Coen
Site Admin
Posts: 1124
Joined: Fri Mar 26, 2004 7:00 pm
Location: Amsterdam, Netherlands
Contact:

Re: ActiView exes hang

Post by Coen »

That's useful information. If ActiView Light works, then we can be sure that the driver itself, as well as the communication between ActiView/LabVIEW and the driver are all OK. The basic acquisition (incoming data stored in ringbuffer by driver, and data read from the ringbuffer by ActiView) apparently work without any problem. I am sure that Synctest will not show any problems either.

It seems to be a pure LabVIEW bug, like a problem with the chart display (video processing ?) or switching of tab-pages (switch between charts). Did you only see errors with the acquisition running, or also when manipulating controls with the acquisition stopped ? If the errors only occur with the acquisition running, then I would suspect a chart display bug. Do the errors also occur when the "about ActiView" tab-page is selected with running acquisition (and no scrolling charts are displayed) ?

It's puzzling that the problem only surfaces on the target computer, and not on the nearly identical development computer.

Best regards, Coen

msergi
Posts: 5
Joined: Tue Sep 16, 2014 6:49 pm

Re: ActiView exes hang

Post by msergi »

Synctest appears to have an issue... below is the output. I had the same results from two different computers, one of which has not had any issues with ActiView. I tried increasing the ring buffer size as well and had the same result.

****************************************
Parameters can be, in any order:
Time to run, in seconds:
default is 600 seconds (10 minutes)
0=unlimited

Ring buffer size, in 512 byte units,
must have a 'b' or 'B' prefix:
default is B65536 for 32 MBytes

Ring buffer extra bytes beyond the 512 multiple,
must have a '+' prefix and be a multiple of 4:
default is +0 for no extra bytes


Hit any key to quit
****************************************
sizeof long = 4
ring buffer size = 33554432
at second 0, buffer 0, #bytes 131072, seam 131072, word throughput/channel/sec 0
number of channels = 82
missing sync 1: buffer 0, [164] = c00f6c00
sync found again at [172] = ffffff00
missing sync 2: buffer 0, [254] = c00f6600
sync found again at [262] = ffffff00
missing sync 3: buffer 0, [344] = c00f8200
sync found again at [352] = ffffff00
missing sync 4: buffer 0, [434] = c00f9700
sync found again at [442] = ffffff00
missing sync 5: buffer 0, [524] = c00f8b00
sync found again at [532] = ffffff00
missing sync 6: buffer 0, [614] = c00f9300
more than 5 missing syncs in 1 buffer
stopping ...
Aborted, hit a key to exit
Done!!!

msergi
Posts: 5
Joined: Tue Sep 16, 2014 6:49 pm

Re: ActiView exes hang

Post by msergi »

With regards to your other questions, it almost always occurs just after clicking the Start button. If I have the About tab selected it automatically switches back to the chart tab with the downloaded vi. However in our modified version the default tab is not one of the chart tabs so that may rule the drawing on the charts... the tab we show has some text entry controls for other test information. The small sync graph at the bottom is still displayed though, I can see if I can disable that perhaps.

I installed the full LabView IDE on our target machine in case it was something odd with the exe or run-time engine. However the hangs and crashes still occur. Interestingly when the access violations occur LabView completely crashes, not just the VI. I had been hoping that running from within LabView would give me more debugging information!

This morning I also tested with a different computer that runs Windows 7. I tried our LV2013 compiled exes on that computer and ran into no issues, despite repeated restarting of communications and the software. So whatever error is occuring seems to be specific to the hardware (or sw configuration) of the Surface Pro 3 we are using.

Best regards,
Michael

Coen
Site Admin
Posts: 1124
Joined: Fri Mar 26, 2004 7:00 pm
Location: Amsterdam, Netherlands
Contact:

Re: ActiView exes hang

Post by Coen »

The Synctest errors are just because this is an older version of Synctest that cannot yet differentiate between sync words and empty channels. Sorry for the confusion, we will put the current version of Synctest on our website soon.

Anyway, your Syntest run confirms again that the USB driver is working OK.

So it seems an incompatibility between some LabVIEW function (both in 8.6.1 and 2013) and the Pro 3. You could try to find the cause by disabling sections of ActiView code until the error disappears. If the piece of LabVIEW code that causes the problem is known, then I can try to come up with a workaround. However, most probably we will have to wait until National Instruments fixes the bug in a new version of LabVIEW.

Best regards, Coen (BioSemi)

Post Reply