Ben,
In this case you would want all the code in the loop. What would happen if you didn't have your frequency and amplitude controls inside the loop is they would be fed from outside of the loop and you couldn't change them once it was started. Putting controls and indicators inside of loops means they are read or updated on every loop iteration. This can be very processor intensive, but for your beginning example it's fine.
The way the express DAQ Assistant works is that it eliminates the need to configure all the other DAQmx functions like start task, write/read, stop task. It's meant for quickly setting up and acquiring/generating signals. Alternatively, all the settings within the DAQ Assistant can be configured using the other functions on the DAQmx pallette. More developed LabVIEW applications probably rarely use the DAQ Assistant, but for this kind of simple stuff it's easy and works well.
The Wait function is controlling how often your loop iterates. In the edited VI attached, it's happening every 10 ms. If you don't have a Wait function in your loop it will run as quickly as it can. Try it and watch your CPU chart in task manager go to 100%. When choosing how often a loop should iterate, the general rule is to make it run no faster than it has to.
Another note about loop timing: when you set up a loop for acquisition, the loop speed can be dicated by the data acquisition depending on how you have the acquisition set up. If you were doing data acquisition that was limiting your loop speed to 100 ms and you put a Wait function in your loop set to 1 ms, the loop would only iterate every 100 ms. The Wait function wouldn't be doing anything in that case. It's like a governor that tells the loop to never iterate faster than the time you wire to its input.
Eric
