LabVIEW
change value during while loop of subvi
Hi dabrainz,
rethink your program design. THINK DATAFLOW…
- How should we debug using just some picture?
- Why do you write to a value property node (of "Indicator") to just read the very same value back from a value property node? Why not use a simple wire?
- I don't see any of your switches in the timer block diagram…
- Why do you use property nodes where a simple wire would also do ?
- What happens in your "Soap application" loop when the timer outputs a FALSE? Why is there no wait in the loop? What's the purpose of the special output tunnel mode?
To send values to subVIs you should prefer wires. When the subVI will block your mainVI you should either have no loops in the subVI or run it in parallel to your main loop.
Hi dabrainz,
It's a condition-tunnel. So it runs as soon as the timer gives true. It waits for the true to come while it's false.
Wrong.
It will not wait for any TRUE. It will either stop immediately or it will run forever (burning the CPU). THINK DATAFLOW!
I'm limited to LV2011 here right now, so you will have to wait until I get access to a different computer with LV2014 installed on…
I had a brief glance at your code, and, as Gerd already said, it shows a complete lack of understanding of dataflow. The small inner while loops in the toplevel program will either run forever without any possibility of stopping or they will stop immediately, depending on the value of the boolean when the loop starts. Once the loop starts the value of the boolean in the tunnel is fixed and will never change. Since the subVI can only exit when the time has elapsed, the boolean is always true and the inner loops can be deleted without any change on program functionality. In fact the boolean output of the subVI is redundant information that is not even needed.
Now let's talk about the subVI:
- Why is there a hidden stop button as an indicator? Since it cannot be operated and there is no programmatic access, it serves no function and can be deleted.
- Why do you use value property nodes to set the elapsed time and the boolean? Since these indicators will get valid values before the subVI completes, these actions are not needed.
- Why do you constantly re-read the cluster with each interation? Since these values are recevied at the subVI call, they can never change during the current call.
- Why do you do all these fancy computations with start times, etc. All you need is the elpased time express VI. It will do all that internally for you already.
- Why do you reset the elapsed time express VI with every iteration of the while loop? You should only reset it on the first iteration (do an =0 on the iteration terminal, for example).
- You only need one instance of the elapsed time express VI. The instance before the loop is not needed.
- You don't need to wire the error across the while loop, since it is loop invariant.
- Overall, you have way too much and way overcomplicated code.
I only glanced at the code for a few minutes, so there are probably way more monsters crawling everywhere. 😄
You should really re-architect this as a simple state machine. You don't even need the subVI. Have a look at the design pattern templates that ship with LabVIEW. Also have a look at the solutions posted in the certification forum as well as the review comments.
At this point you are not ready for the CLD, but once you gain more experience, that day will come. As exercises, I would suggest to write very small programs that only demonstrate one aspect of what you are trying to do. Then explore it with debugging and execution highlighting to fully understand how dataflow works. Once you have a clear understanding of each aspect, combining the parts into a functional bigger project will be simple.
