Chart Reseting on Host

[Milqman]

I am running an RT system and have the Host.vi on my computer (same subnet).  When I run the Host.vi I my chart appears to be reseting (starts blank, begins filling with points, then goes blank again and begins filling with points again).

Below is my chart.


I am sampling at 1kHz inside a loop that cycles every 10ms (there are 10 waveform terms per loop).  Is this a buffer problem?  Am I doing something else wrong?  The chart gets to this state then blanks, so on and so forth.  The period of the blanking is not constant.  I have tried a normal buffer, and a FIFO buffer with varying settings but I cannot seem to eliminate this behavior.  It appear that the bigger I make the buffer, and the more deterministic I make the application, the worse this behavior gets.

I am sure there is probably a simple answer for it.  You guys are always on top of things, so thanks in advance!

~milq

[Ben]

I have seen this behaviour myself with charts when using waveform data types.

If you are not using WF datatypes, you can ignore the rest of this post.

What was happening with me was that my code was presenting inconsistant data to the chart. What I mean by this is that new updates to the chart where over-lapping the time period already displayed on the chart. Here is an example of how this happens.

1) You present a set of data to the chart that starts at 12:00 has 60 samples and the dt is 1 second. The end of this data then should end at 12:01.

2) Another update occurs that specifles the T0 of 12:00:59. Since the new dat over-laps the old data in time, the old data is tossed and the new data is displayed.

I am not sure if this is what is happening to you but it does fit with your description.

Trying to help,

Ben

[Milqman]

Ok, I will look for that.  Not sure how I will get rid of it, but I think I can figure out whether or not it is the same problem.  I guess tweaking the period on my loop might fix that.

Thanks!

You guys are the best!

[Milqman]

That's not it.  I set it up to sample twice per loop (where the period of the loop is much much slower than the period of the sampling).  The issue was the same.  I did find another interesting issue though.  The graph works just fine when I display it on the target, no blanking at all.  So my thoughts would lead me to my network (which is very very fast) or some type of buffering issue (which I know very little about).  I am going to experiment with a crossover cable and see if that fixes anything, if it does, then its the network.  Otherwise, it is a labview issue.

Does anybody else have any ideas?

Thanks!

~milq

[Milqman]

Anybody else have any ideas?  I am stuck on this one.  I want to see about 5 seconds of graph and this thing is to the point where it is reseting every .5 sec or so.  When I put a graph on the target's VI it looks just fine, somewhere between there and the graph on the host side, there is a data problem.  It makes me sad!

Any other suggestions would be very welcome,
~milq

[Ben]

"Anybody else have any ideas?  "

Post an example and share what version of LV you are using.

This is also similar to a bug I saw in the web-server that was present if I used property node >>> value to do the update vs using a terminal (or was it a local).

I think that was LV 6.1.

That is all I can offer for now.

Ben

[Milqman]

I would have come out with more details but I sort of assumed that it was a common problem because it occurs everytime (even when I build something simple from scratch).

I am running LabVIEW 8 and the Real-Time Module

Here is the VI on the target, the issue here is towards the top, the circuit chunk ending with the "Analog" shared variable.


Here is the host VI



Again, trace "Analog" and you can see where the graphs in question are.

With the troubleshooting I have done, the graph looks fine on the target side and not fine on the host side.  My next step will be to put the graph of the variable in the other loop, and see what it looks like there.  At this point I am considering the network as a contributor, and the buffering.

If there is anything else you want to see, just ask, I have been banging my head against this for too long 😞

Thanks!
~milq