Error 116 & Warning 10694

[golubovski]

Hello colleagues,

I know that the "error #116" issue has been discussed a lot lately, but none of the posts helps me cmprehend the source of my problem described below, epecially given that my concept had worked well with previous versions of LV.

I use "Sync=4472 Trig=6713.vi' to synhcronize 72x AI (9x PXI-4472) with 2xAO (PXI-6713) on a shaking table. This VI successfuly saves sequences of 3D array in a binary file. The parrent demo VI is "write.vi" which performes the DAQ and streaming to disk. No porblem here, except for the warning #10694 which I also do not understand where it originates from. The missing VI is a 2D array constant containing 10 minutes of signal generatuion sampled by 1000 S/s - I just subset the length I need according the acquisition duration.

The error #116 is reported by the "Read from Binary File.vi" in the demo "read.vi". I do not see what causes this error since the data type definition I wire to the "Read from Binary File.vi" is the same 3D array of doubles!?!

Can somebody "116" experienced take a look at the this simple and clear demo and possible identify my mistake?

Thanks in advance

[smercurio_fc]

As far as I can tell, you're writing a 2D array, not a 3D array. Also, your "Sync=4472 Trig=6713.vi' is password protected. What's the point of uploading a VI for others to look at if it's password protected?

[golubovski]

I aplogize for the password, I forgot. It is put there to disable unauthorized access to the DAQ part to prevent the PXI as well as the shaking table from damaging! the password is:

1281940

As for the 2D//3D issue - I am sure it is 3D:

1st D = device

2nd D = row (channel)

3rd D = column (data)

any comment

[smercurio_fc]

OK, now I see where you're writing the 3D data. Off-hand I can't see anything immediately wrong. Since we wouldn't have your hardware setup, we would not be able to run your VI. Can you upload a sample data file that gets written?

[golubovski]

Hi again and thank you for your time. Uploaded are couple of files, both containing sampled data of 1 second at 1000 S/s. The only difference is the chunking which shoud prove correct data streaming to disk - the "10" file uses chunks of 10 samples per channel to read and stream, the "100" file uses chunks of size 100 samples per channel.

The files are not of TXT format! I had to put that extension for this forum to accept their upload. So you can safely remove the extension.

I use transpose matrix for the chunks (72x10 / 72x100) so upon successful data retreival from the files you should get straight 72x1000 samples (1 second acquisition at 1000 S/s)

Thanks in advance,

[golubovski]

One more strange thing. LV help says [double] is 8 bytes, yet straming stores 4 bytes per sample?!?

Any reasonable explanation?

[smercurio_fc]

In your code you're using SGL, not DBL. SGL is 4 bytes. DBL is 8 bytes.

The problem is that you're not prepending the array size, so LabVIEW is not interpreting the data properly. In the VI where you're writing the data you have a constant of False wired to the "prepend array or string size". Change it to True. Then your read should work. Sorry I missed this constant on my first look.

[smercurio_fc]

FYI: As a follow-up, you can take a look at this article for more info. It's for 2D arrays, but the same principle holds for 3D arrays.

[golubovski]

smercurio,

Thank you, that does the job. I must have changed prepend to off by mistake. I have just tested it on the concole PC via remote desktop.

Now I have another challenge - the "Read from Binary File.vi" needs to know the number of chunks in advance and produces an array of chunks. I guess I should put this information at the beggining of the file?

[golubovski]

I see you have anticipated my last posting.

Thanks again,