LabVIEW

cancel
Showing results for 
Search instead for 
Did you mean: 

Read From clones to array

Solved!
Go to solution

Objective:
I want to create multiple clones of an FGV and put their data into an array on my front panel when called. 



I am pretty sure I have the creation of clones correct: I check a strictly typed vi ref and ID cluster for the desired FGV ID, if it exists, I pull the vi ref and send it, if not this code runs:

 

Make New CloneMake New Clone

When probing the [Chamber No, vi ref] cluster, I can see each time a new chamber number is added, the array increases in size.


Here, I'm reading from the vi ref (passed to a different vi through a shared queue). Ideally, it should check the vi against the ViRefArray cluster to see if it should add a new element to the 1D data array or replace an already existing element. 

Call CloneCall Clone

Unfortunately, even though the id's are different, LabVIEW is returning "found" for the ViRefArray. this is making my clones overwrite each other rather than be added together.

 

Capture.PNG

 

0 Kudos
Message 1 of 12
(4,148 Views)

You're making things really complicated. Not sure what you want, but I can speculate about some problems you'll run into.

 

So your keeping an array of references. These are just numbers. If the VI stops, or when it's not even run, the reference will become invalid. The number will still be in your array.

 

Do you start the VI somewhere else? Doesn't show in your code.

 

Looking at those update times of the probes, I'd think your test is just off. I'm pretty sure the Search 1D Array is bug free.

0 Kudos
Message 2 of 12
(4,085 Views)
Solution
Accepted by ATE-EGNE

Two issues

 

  • When comparing references you compare if the thing it points to is the same.  Every ref in your array is an instance of the vi interface type 0.  FIX:  Add Property Clone Name to the cluster and compare that instead.
  • Your options flag os 0x120.  You want 0x160

"Should be" isn't "Is" -Jay
Message 3 of 12
(4,061 Views)

@JÞB wrote:

When comparing references you compare if the thing it points to is the same. 


Didn't know that. Thanks!

0 Kudos
Message 4 of 12
(4,048 Views)

After reflecting about this for a while, I think this might be a relative new feature (LV10? maybe 8?). So before that could it be that the actual numeric values where compared?

0 Kudos
Message 5 of 12
(4,042 Views)

@JÞB wrote:

Two issues

 

  • When comparing references you compare if the thing it points to is the same.  Every ref in your array is an instance of the vi interface type 0.  FIX:  Add Property Clone Name to the cluster and compare that instead.
  • Your options flag os 0x120.  You want 0x160

Have a kudo from me too. I had never run into this problem before, and now I never will.

Message 6 of 12
(4,035 Views)

wiebe@CARYA wrote:

After reflecting about this for a while, I think this might be a relative new feature (LV10? maybe 8?). So before that could it be that the actual numeric values where compared?


Nope, allway was.  Type cast to I32 will compare the value.


"Should be" isn't "Is" -Jay
0 Kudos
Message 7 of 12
(4,018 Views)

But stepping back to the OP though (and getting OT again): why is this even needed? Seems to be easily avoidable? Why the variant? Seems like trying to kill a mosquito with a bazooka to me.

0 Kudos
Message 8 of 12
(4,011 Views)

wiebe@CARYA wrote:

But stepping back to the OP though (and getting OT again): why is this even needed? Seems to be easily avoidable? Why the variant? Seems like trying to kill a mosquito with a bazooka to me.


I wanted to do it this way for two reasons:

1) Proof of concept (also wanted to learn more about passing clones/vis)

2) I wanted to separate out processing the data received by the computer so that receiving data (from one COM) would not be held up by processing the data. Also, if one clone starts after another, but finishes before, it will be able to be displayed as soon as it's finished.


As for the Variant, is it better to not use one? Most the P/C states I see pass an enum and variant cluster. Is there a reason to avoid doing so?

0 Kudos
Message 9 of 12
(4,004 Views)

I'd not recommend variants when they can't be avoided. It's not always possible, and I can't be sure from the images if they are needed or not.

 

Both images are "complex". Not impossibly difficult, but too complex for my taste, knowing only a bit about what it does. It's hard to see at a glance what it's intended to do. That's OK for a few VI's in a project.

 

When starting dynamic VI's per chamber, I'd make a Chambers.class with an array of VI references and an array of chambers in it's private data. It would be very easy to see if a chamber is started, add a chamber, or invoke a method on a chamber, since all needed information is nicely kept in the class wire. The chambers will still be running dynamically, but the user of the class (you) doesn't need to know that.

0 Kudos
Message 10 of 12
(3,998 Views)