creating a matrix of arrays

[RavensFan]

No.  You don't turn the 3rd index into an array.  That makes no sense.  Remove that "Build Array" function from the middle of the broken wire and turn that 0 into some other value such as 5 or 10, or whatever number of pages of data do you need.

What you want is a 3-D array.  It consists of rows, columns, and pages.  If you indexed out a page, now you have a 2-D array of rows and columns.  If you indexed out a page and a row, you'd have a 1-D array of columns.

You have to start visualizing your data in a 3-D cube and depending on whether you are specify something in one dimension only, your result is a 2-D array.  If you specify in two dimension, you are left with a 1-D array.  If you specify in all 3 dimensions, then you are left with a single element of the array.

[krivan]

Yes, it is clear but the reason why I tried to put there the array as a third dimension is that I don't know the exact size of the "3-D cube". I cannot pre-specify it. I namely know how many sensor I have [this is the 2D array because the sensors are arranged into a matrix] but I don't know how many thousands of measurement points I need to acquire. These measurement points would make the third dimension, I would put them into an array that is constantly growing as the data acquisition is going on. When the cube reaches a certain limit, it should be passed to another parallel running while loop that streams the content of the cube into the respective files.

Probably there is a much better solution for this problem but this is the way how I thought about it.

So there is a 2D array, a matrix. To each element of the matrix connects an empty array that would make the third dimension. And these attached arrays will continuously grow as the measurement data are saved into them.

[RavensFan]

First, a 2-D array is not a matrix.  They are actually 2 different entities with in LabVIEW.  However, there are a lot of similarities between them and many of the array functions

You can grow the 3-D array as needed.  You can start off with a 5 x 5 x 0 array if you wanted, (5 x 5 "matrix" of sensors and 0 data points collected) then use build array to put on a 5 x 5 array so you now have a 5 x 5 x 1 array, and so on.

[krivan]

Brilliant idea, thanks! I'll give it a go very soon.

However as I started to think about the application I ran into the question of how to stream out such an amount of data. In reality there is a 20x20 sensor matrix, the PXIe-6358 will collect say 100 data points of each sensor in a complete sensor matrix readout cycle. The frequency of the complete sensor matrix readout should be at around 1 kHz. This will repeat until the measurement stops that takes in worst case days.

I thought of a kind of producer/customer architecture like solution. Namely, after every 1000 data acquisition cycle the VI would start to store data into a new 3D cube and would send the old 3D cube to a separate while loop that would stream the data into files by TDMS. 

Would it make sense to do it with producer/customer queues or the amount of data would be too large for this? Any ideas are welcome.

[Ned_Kelleher]

I need some help! I am writing code for a force plate - FX,FY,FZ + moment X, moment Y and momentZ

So I have a 6x6 calibration matrix for my force plate

I am acquiring data continuously but my program will wait for a wireless pulse from a Delsys electromyography program, then collect and write to a lvm file for 5-10 seconds. My question, for all 6 channels or columns in my lvm file seem to display a cyclical pattern. I do not think I am getting the numbers I want.

I will attach vi

I have captured 5 seconds of data standing on the plate; the values for Fz are in the 3rd column (FZ).

Lvm format Fx,Fy,Fz, torqueX, torqueY, torqueZ

[rex1030]

Ned, start a new thread for your question and you will get help quickly.  (No thread hijacking lol) Post your VI and a screenshot in your new thread when you ask your question