Synchronize two while loop's start

[Joaquinh]

I have two sensors sending data at different speeds:

- The first sensor, A, is what I based my VI from. It is a while loop that uses event and case structure to toggle between: idle, start communication, start measurement, stop measurement, stop communication. 

- The second sensor, B, is an serial communication running in a while loop. (It is at the bottom of the VI attached)

I need to have both loops running at the same time but I need the second sensor B to start when the first sensor's event toggles to "START MEASUREMENT". Both have to receive message at the same time, as the time stamp is crucial.

I have attempted to 1) place them both in a while loop but the time difference created a loss in packets. 2) local variables, but "boolean latch is incompatible with local variables", 3) have a wire direct from the event of sensor A to a event of sensor B but it did not start the second loop.

I have looked at notifiers, queues, consumer/producer --- without any result. 

Any guidance would be great!

[Kevin_Price · Accepted solution]

---

@Joaquinh

---

I have looked at notifiers, queues, consumer/producer --- without any result. 

Well, then you probably better look harder.  That's where the solution's gonna be.

Your serial comm loop needs to be off on its own because you can't force timing sync with a serial-connected device.  The best you can do is to associate a timestamp with the data it produces.

Notifiers and queues (plus user-defined events and the more recently introduced Channels) represent the accepted best practice methods to send signaling (trigger-like) and data between parallel loops.

You can't really force both measurements to occur at the same time.  You can kinda make sure you send the query string via Serial Write a.s.a.p., but you have no control over how quickly your instrument will respond.  Even when it does, you may not be able to know what point in time that piece of data will represent.  Does the query initiate a new measurement?  Or is the instrument reporting the most recent old measurement it happened to take?  Either way, there'll be some notable uncertainty in your ability to tag the data with an accurate timestamp.

I can't quite tell what's going on with sensor A, but it kinda looks like it might also involve communication that can't be controlled in time (USB and Ethernet?).  If so, you'll just have to try to associate a *separate* timestamp for the data you get there.  Some of the same issues would likely apply -- the best timestamp you can come up with will still have some uncertainty error.

Since you're writing to TDMS, I'd write these timestamps as channel data with names like time_A and time_B.  Each channel of sensor data gets its own associated channel of timestamp data.

-Kevin P