Longest charting history

[NIquist]

Sure, just make your chart history large enough and that's what you get by default since the x-axis will be set to autoscale.  If you do want to zoom in you have to turn off autoscaling and then reset your x-axis range manually (then if you turn on the X Scrollbar you can slide the data around easily).  Keep in mind that the data you see is compressed by some averaging routine so in the big picture you may not see tiny details in your data like spikes or glitches.

[JÞB]

---

@labviewleroy wrote:

so, is there a way to ONLY see the big picture? I dont carea bout zooming in to finer and finer details

---

Yes,  The KB article I linked in my previous post offers several suggestions on how to decimate large arrays to display data on charts and graphs without copying the entire large array. The chart terminal DOES make a copy of the data on the BD wire so, that's 2.7Mpts x 2 x Bits per point (a DBL is 64bits). 

[Leif Dams]

Hi,

I have a similar application where I will run a so called life time test of a system. That test will last for over 70 days. It is obvious that both the PC and Labview must be stable and run for that period. Or else it will be useless.

I don't know the nature of what you are measuring, but it is no point of doing one measurement every second if the signal changes at a much slower frequency.

I am doing one measurement every 10 seconds, and save it on file. To have a informative presentation I use one strip waveform chart for the last few hours. For the whole test period I use a scope waveform chart showing only every 60th (every 10min) measurement. I am not very happy with this solution. If anybody has an idea of how this can be done more nicely I would appreciate that.

[Joseph_Loo]

My length is about 10 days. I collect data at every 5 seconds. This data is kept in multiple files, I create a new file every hour. I then create another process file that is good for the whole run. This file contains date every 60 seconds. This allows me to get an overal view but if I need to look at things more closely, at the data, it is available.

[johnsold]

What is it about that which makes you unhappy?  You were not very specific so it makes it more difficult to offer suggestions.  Good questions are much more likely to generate good answers. 

Any chart or graph (using a line plot) ultimately can only display a number of points less than the widht of the plot area in pixels.  If you feed it more data than that, LV will reduce the data to the number of pixels.  The LV help (at a quick glance) does not describe the algorithm(s) used for this data reduction.

Your data rate (1 measurement every 10 seconds for 70 days) results in more than 600000 data points, far more than can be displayed in one graph.  Dividing by 60 still produces more points than will fit on one graph.

What do you want to display on your graph or chart? Do you want the user to be able to manipulate the display? What options should the user have?

Lynn

[Ben]

And for another approach...

You could use LV DSC to do the logging. i have 5-6 fume hoods used by the CDC that have been up and logging  almost ( I did have to take them down to update them from LV 6 to LV 8.2) continually since 2001.

DSC will let you do historical queries on whatever time period you want. It also offers arciving functionality and the Historical Trend Viewer in MAX will let you look at any value from any of your archieves.

Ben

[viScience]

I second Ben's approach using the DSC and the Citadel Database.  I have been logging ~ 200 channels at 20 Hz for the past 2 years without a hiccup.  As Ben also pointed out the Hypergraph activeX controls allow you to create your own viewer in LabVIEW or you can use the Distributed System Manager to view any real-time or historical data. (I personally like to use the DSM rather than MAX)