Jay - what you said sounds really useful, however it's a little beyond me knowledge - I read up a little on DVRs, and it sounds promising, but it will take me a while to figure out how to convert my code. Specifically, I don't really see how to wire a DVR to the Mathscript nodes.
I tried to inline many of the "data chugging" vi's, however even here I run into a maze of trying to figure out whether inlining is better than subroutines, since those methods are more often used for speed rather than memory management (if I understand correctly). Also, by inlining do you mean selecting the "Inline vi" option, or rather using the IPE structure? What's the difference for a simple vi performing arithmetic operations?
altenbach - The reason I blamed the conditionals is because the problem started popping up more often after I added them, but you're right that they might be innocent. I'm sorry about the spaghetti / hairball, I tried to make it rather clean, but I didn't want to bundle too many wires when I knew they will soon split into many directions. The conditional terminal is in the top left part, connected to timing shutoff. Local and Global variables mostly pass physical constants, that have been calculated prior to the main for loop so there is no race.
I don't know if I'll be able to reduce it to a simple sim without losing the character of the vi. If I simplify too much, it might not reflect the original problem. I'm trying to work with a memory debugging tool, we'll see if I get anywhere. Surprisingly, after a PC reboot I was able to record much longer datasets than before without getting that error... not sure why.
Thanks for all your help so far, and sorry for the long reply time.
