What makes chart go wonky?

[Ben]

XP

[JeffL]

Hi everyone,

I'm currently talking to R&D to see what we can do.

[Vivek_N]

Hi,

I am not sure what exactly is causing this behaivor. Does minimizing and maximizing the front panel, or  moving the window around or off the screen fix the chart display? Also, have you tried changing the chart type (sweep, strip, scope) on the fly? Does this make a difference?

As I am not able to reproduce it on this end,  I would like to use the VI you are running. To simulate your hardware setup here, you could export your MAX configuration for me to load on  a PC here. To troubleshoot this problem, we will need to reproduce it on our end, so the VI would be very helpful. 

[CoastalMaineBird]

I cannot give you the stuff to do it, as I have a 950+ VI program running, with PXI hardware, I do not use MAX, and even I can't reproduce it on demand.

I do know that the part where the plot lines were misdrawn (in post #1 on this thread) was cured by reducing the workload - I was  opening-seeking end - writing a line-closing a 4 Mbyte file multiple times per second. 

When I applied some sense to that, that particular aspect went away, and the CPU usage went from 60% to 2-3 %.

The business with the gridlines being incorrectly drawn is also random.  I cannot reproduce it on demand.  I suspect it's instantaneous CPU load, and doesn't show up in the CPU meters, but that's gut feel, not hard facts.

When it happens, it's always when there's a transition of some kind - that last pic was the end of a test, which meant the tab was programatically switched, all sorts of events happened, etc.  The chart has TWO scales, both set to autoscaling. Notice that the left scale (800 / 780 / 760, etc) does NOT agree with the yellow gridlines until the test concluded (11:45:55).

It DID agree during the test however.  Something cause the scale to change, and the  numbers changed, and the NEW lines changed, but the old gridlines did not.

[CoastalMaineBird]

Does minimizing and maximizing the front panel, or  moving the window around or off the screen fix the chart display? Also, have you tried changing the chart type (sweep, strip, scope) on the fly? Does this make a difference?

I haven't tried changing the chart type, as I don't enable that menu for the user, but a lot of things will "fix" the display. 

Changing the Ymax via mouse will do it, moving the window COMPLETELY  off will do it, it will fix itself in a few seconds. 

[CoastalMaineBird]

CORRECTION:

It just happened again (the part with the confused gridlines, not the part with the confused plot lines).

I did nothing but change a channel, which just chooses which element of the data stream array to plot.

Changing the channel on one chart gorched the scales on BOTH charts.

REPEAT:  There are two selectors for the top chart and two selectors for the bottom chart (four channels in all). Changing ONE selector screwed up the scales on BOTH charts.

CPU usage is sitting at 2-3 %.

The channel change caused a scale change (on only one plot).

Dragging the window offscreen and back on DID NOT straighten it out.

Pulling down the menu in front of the chart DID NOT straighten it out.

Switching tabs away from the charts and then back to them DID straighten it out. 

[CoastalMaineBird]

Here's another case that just happened.

The host processor was NOT especially busy at this time.

I was under the table at the time, moving a monitor connection from PXI box to another computer (unrelated to this test). 

When I popped my head back up- the display was like this.

It took it 20+ seconds to right itself, during which it jumped from one wonky display to another, but I was only able to capture this one shot.

Again, there was NO special CPU activity, no windows flipping, nothing to cause a CPU usage spike.

[JeffL]

Hi again Coastal,

Obviously there is some problem here, but unfortunately none of us have been able to recreate it on our end. If you are able to create a VI that displays the issue at least intermittently, it would allow us to get to the root of the problem.

[CoastalMaineBird]

Well, this VI doesn't gorch up the plots every time, but it WILL gorch up the grid lines - perhaps that's related.

 

I'm attempting to add a feature - I want to switch (at run time) between having two plots on two independent scales (easy enough) and having two plots on the same scale.

It's easy to set Plot #1 to connect to scale #0 instead of scale #1, based on the switch.

But I ALSO want to have the scale on the right show the same numbers as on the left (including any autoscale changes).  Since it doesn't have a plot to play with, I copy, every display cycle, the scale 0.RANGE to the scale 1.Range.

That seems to work, but it leads to the wonky gridline every time.

Instructions:

Run the attached VI for at least 60 seconds (to let chart fill up).  After it starts scrolling, flip the SAME SCALES switch UP, then DOWN.  After the DOWN, the gridlines go wonky.

[johnsold]

Those gridlines are correct for the left scale when the scales are different.

However, it appears that (1) the gridlines for the left scale do not change at the same time that the scale changes and (2) the entire (visible) plot changes scale when the switch is flipped but the gridlines do not (see(1)).  I also noticed some inconsistencies, which I have not attempted to characterize, which may be related to autoscaling.

I ran it on the Mac version of LV2009.

Lynn