Export subset data

[RJJE]

Hi David

Sorry, I didn't save the project with the added post processing. The following screen shots show my methods.

NI_1 shows the reopened project done about a week ago with the first log as the active one. In my situation I have to record data first, look at after and then decide what data is needed and post process. So the red & white traces (motor current) from about 10:15 to 10:20 is the data I would want to snip out for post process.

NI_2 shows the playback mode and the settings I did there.

NI_3 shows the step added for playback. Scaled signal has been right clicked for 'enable recording'. This morning I noticed there is no trace in the scaled window.

NI_4 shows the results after hitting Run in playback mode. The new log is created but nothing is shown on the Data View screen when the log is dragged and dropped on to it.

This morning I had some time to try this. Close and reopen the project without saving. Follow steps as done before with the exception of changing nothing in the "playback options" tab.

NI_5 shows correct operation of the scale step, trace result in bottom window. Hit Run and the system plays back the original log.

NI_6 shows the result, I didn't let the whole log playback. New log was created and the data is displayed. I can't explain why the red & white traces were not included as I did not change the input signals of the step which is set to ai0, ai1, ai2 and ai3.

I have 2 NI 9215 cards and a NI 9421 card in my CompacDaq rack and will test them later on today but I think they are fine and it is my lack of experience using SE.

Thanks

[RJJE]

Only 3 attachments per reply. Heres the other 3 screen shots.

[RJJE]

David

This may be of interest. If you leave the start/stop play back time markers at the beginning & the end the Scale step will function as I mentioned above.

If you move the start marker off the beginning the step will not generate an output and will not show the step result trace in scaled signal result graph.

In NI_7, I started with the starter marker on zero and both traces in the step setup window showed the first 50 msec of the trace. NI_7 shows the result after I moved the start marker off zero. The top trace shows starting at the new start marker position whereas the bottom still shows the first 50 ms of the trace.

Don't know if this means anything.

Randy

[RJJE]

Did it, but not sure how.

Playing around with the 'Playback Options' tab and its settings. I would change a setting then check the 'Step Setup' graphs.

Somewhere along the line I ended up with the full trace time in the 'input signal' and 'processed signal' graphs. I then deleted the scaling step and added the subset and sample step. The 'Step Setup' graphs remained the same. See NI_8 screen shot. Here I adjusted the start/stop markers, set the subset signal to enable record and hit Run.

A new log was created and viewing it gave me my time snip but missing 2 of the signals. See NI_9.

So close.

[David L.]

Hi

Thank you for the replies and detailed information – great screenshots! I believe what may be happening is that occasionally a full block size (samples) is not taken, and therefore not logged because the 'Ignore Last Iteration if Partial' box is checked. Does the same result happen if you make your block size smaller? (For result purposes, please try a range of numbers – 100, 1000, 50,000, 200,000, as to eliminate false negatives by not running enough tests). Also, I would suggest running some test with the ‘Ignore Last Iteration if Partial’ box unchecked – opposite of what is seen in ‘NI_2’ from your previous post.

Regards,

Message Edited by David L. on 10-21-2007 08:41 PM

[RJJE]

Hi David

I will try further testing, but I don't fully understand how 'sample block' & 'iterations' etc fit in the picture.

Is there a detailed explanation of their functions?

Randy

[David L.]

Hi Randy,

The direct definitions for the ‘block size’ and ‘ignore last iteration if partial’ can be found in the help file (Open SignalExpress >> Help >> LabVIEW SignalExpress Help – see picture below). Basically, the block size is how many samples at a time the file will be cut into. The ‘Ignore last iteration if partial’ will setup SignalExpress so that if the last iteration is not a full iteration (does not have enough samples for an entire block size) it will not be processed. This is why we thought you may be seeing ‘blank’ samples.

I would like to take a step back to a slightly previous post, where it seemed like the issue was resolved, except that 2 signals were missing. Is this still the case that we can pursue? I would like to get a more clear definition on this, and therefore ask for another picture, line NI_9, but showing the legend as seen below. This will let us know how many channels are being graphed, and if a channel that you are expecting is not being included (This is what I interpret by 'missing 2 signals' - 2 channels are not being graphed). I would also like to ask for your latest SE project, along with the data files so I can see what is missing. The files are stored here: C:\Documents and Settings\UserName (All Users, administrator, etc)\My Documents\SignalExpress Data.

Regards,

Message Edited by David L. on 10-25-2007 07:25 PM

[RJJE]

Hi David

I can't seem to re-create NI_9 screen shot. I do remember though, checking what signals were listed and all four were listed but only 2 traces were visible.

I can create sub sets with repeatability now. All I do, is go into playback options and increase the block sample size (just add a zero each time) the return to the subset step and check the x axis time length. I keep increasing the block size until the time window I want is displayed int the subset window. Then adjust the start & stop marker, enable record and I get a new log of that data.

Thanks for all your help.

Randy