Since it sounds like you get 1 point for each axis per loop, you will have to loop the 20 times. Go ahead and autoindex all of the values out. You can use Array Subset to reduce the number of samples to work with.
It really depends on whether the OP wants continuous acquisition or not. I was going on the assumption that he did.
You might want to take a look at the Recent History Buffer Example.vi in examples. You can make separate renamed copies of the buffer vi for each axis, with a different length for each. Old, but useful code. The buffer works as a old style global and can be read out elsewhere in the diagram or another vi.
Old but useful code. The buffer vi can be modified to use your data type. You can creat a separate, rennamed copy for each of your variables with various depths, and they will work as old style globals. You can read them out elseewhere in your diagram or another vi.
