Labview chart with long history buffer

[RavensFan]

How frequently are you collecting data?

If it is once per second, then 7 hours would be 25,200 samples.

But you really should also read this thread.

http://forums.ni.com/t5/LabVIEW/Longest-charting-history/td-p/1673180

[data logging]

i wish i could logg data in to my compuetr as i wish http://forums.ni.com/t5/Multifunction-DAQ/saving-data-into-new-file-in-a-folder-every-time/m-p/16729...

[Xiao@Gemini]

Hello,

I'm running on cRIO NI9035, I would like to set it to 576,000, is the hardware memory capacity enough for me to set up or it will dramatically slow down the process? thank you.

[Michael_78]

Not sure what the ram is on that device, but you could make the datatyepe single precision instead of double precision float. This will halve the memory, I do that in an application with 10 charts each storing 1 sec data for 10hrs. That is on Windows. You could use an i16 to halve again but that would probably be too limiting. Cast the type before writing to the chart if the data is dbl. 

[Xiao@Gemini]

It's running on a compact RIO controller NI9035, it has 1GB memory and 4GB hardrive according to the spec, see the attached pic. I have to use double precision, I'm not sure whether it depends on the hardware or i can just change that number to large one maybe 144,000 and try. Thanks for you reply.

[Michael_78]

It depends on the rest of your application memory use.   You may need to use dbl and save dbl to disk, but why not cast to sgl for the chart? I doubt you need dbl for the chart, unless you are zooming in to tiny ranges you won't have enough pixels to display more than sgl , the chart is just for visual use and should not be used to buffer data for saving or any post processing really.  My charts have 40 traces with large buffer and use about 50MB each. That is quite a big block of contiguous memory. Try it either way anyway, Michael