The fast loop rates are only part of the problem and simply adding a wait is only a bandaid/ducttape solution. You might still run into the same problems under high CPU load conditions, just less often. It's the program design that lends itself to such issues. Fortunately, there is an easy solution in your case. I took the liberty to make a few minor changes that will significantly improve your code.
Since you are using LabVIEW 6, you might not have the event structure, which would simplify things, but here are a few tips:
*** Get rid of sequences and local variables whenever possible! ***They are inefficient, break the dataflow, make the diagram hard to read, or could cause mistakes/race conditions. (mostly all of the above
).
In this particular case, you don't need the sequence and you need exactly one local varable per cluster. (See attached image with suggested code revision).
I wanted to see the Tick subVI, because it was conceivable that it was resetting the cluster. Again, this subVI uses a three-frame sequence and 6(!!) local variables where none(!!) are needed. Again, I have attached an image for alternative code. It is so much easier to read and will run much more efficiently. Enjoy!