The obvious, but not right, solution is to use the "wait" function in LabVIEW. This isn't the right solution because everything inside a LabVIEW loop must complete before the loop can proceed to the next iteration. If you put the wait inside the same loop as you feed the watchdog, the watchdog will get fed once, and then not be fed again until the wait completes and the loop proceeds.
The correct way to do this is store a timer value in a shift register. One implementation: when you first push the button, add your wait time (say, 500ms) to the current millisecond timer value and store that in a shift register. Each time through the loop, see if the current time has reached the value stored in the shift register; if so, your timer expired. Be careful not to use = as the comparison, since it's unlikely that your loop will execute an exact number of milliseconds later.
Someone asked a similar question last year and I posted an example then. You can find it if you dig back through the discussions, but it's a good exercise to try writing your own.
