LabVIEW
How do I create a start/stop button for each separate while loop within my program, when each of them does a different task?
Hi vishal,
I'd suggest starting your own thread with a new question.
Try and indicate what you want to do, what problems you've had so far, and which bits you're needing help with.
I'd also guess the State Machine might be helpful, although you can consider (for monitoring) the use of a Producer/Consumer where the Consumer is your monitoring loop. The producer could be one of the states of the state machine - note that I don't mean that you should nest loops inside one another for the SM/Producer, although you could if it is really necessary. However, a State Machine can certainly manage without that complexity!
If we're looking at the top right corner, then it looks like you're not using Index Array but instead one of the OpenG VIs, perhaps array of variants to a variant, or something like that?
You can just compare (e.g.using equals, Greater than, whatever from the comparison palette) if you want to compare the array size. Indexing will get you either column length or row length as you already described.
In general you can use Variant to Data to convert a variant to a specific type, but I don't think it should be necessary here.
Ok, so let me try break these down a bit:
@vishal231 wrote:
1.I added the index Array block,
but, when I am running this VI , and Pressing the button to log data , Array size block is giving 50,000 samples(i have put 50k as rate and sample per channel in acquistion VI) beacuse of which select is not triggerd and i am not able to go to wait Event again.
The "Array Size" here is the size of a single dequeued element. So it doesn't accumulate over time, as you've already seen.
If what you want is to repeatedly dequeue until the combined amount of data is greater than some amount, I'd suggest using a Shift Register with that value (the total target) being set at the start of acquisition (and reset each time you want to get another block) and then subtracting the size of the current block from the shift register. Check if it is less than zero, in which case you've acquired more than that number of points.
Is that what you need?
@vishal231 wrote:
3. even when I am reading the data from Reader stream from saving VI, and when Monitoring with indicator data Display is not constant it fliashes(Maybe beacuse of delay in getting the data)
I'm not exactly sure what you mean here. Do you mean "waveform graph 2"? Or "Current"? Or perhaps a different indicator?
WG2 will update once for each waveform, but the effect of the way it's structured will be to have most of them flicker through very quickly, then one remain on the graph until the next Dequeue operation, so it will effectively flicker every 500ms probably.
You might want to take a look at placing it outside of the For loop, and passing it an appropriate data format for multiple plots.
If you want them to append to each other, you'd need to use Shift Registers and the Append Waveform (inside the For loop is probably best to append) to store history, but beware that this will also increase your memory usage (because you'd never release previous data).
To store only a certain amount of history, you can search for information about "circular buffers", which essentially forget data once it is old (so for example, retain the last 5 dequeues, or the last 50, but not an infinite quantity). This allows you to know how much memory will be needed (if the dequeues are a constant length). If not, it at least bounds the quantity (and you could implement it in terms of points rather than dequeues to ensure a constant length memory buffer).
Urgh. The quoting interface on these forums is so error prone - I just saw I deleted my section 2 response when trimming quotes before. Oops... I'll try remember what I wrote...
@vishal231 wrote:2.Also, sometimes
"Memory is full" issue Comes up: the top Level VI saving.vi was stopped at Array to spreadsheet string A2T on the block diagram
How to optimize the Memory ?
It seems like you're logging quite a lot of data - does it have to be in a text file? If you can use binary or TDMS instead, that might be helpful for you.
TDMS is a binary file with a specific format specified by NI that is well suited to storing time-series data (i.e. how something changes over time).
If you can use a binary file format instead of the text file you're currently logging, you'll get the following benefits:
- You save a lot of disk space
- You can store more accurate data
- You don't need to convert to strings
- It will probably write and read the file much faster (but I'm not sure how to quantify "much faster" without testing)
If the dataset is just too large even without the string conversions, or if you must use a text file, then you could consider using a smaller data chunk (i.e. dequeue more often, smaller blocks). This will allow your system to write to disk and forget the data faster (assuming you're not keeping it all in memory for the graph as I described in my first response, which should have included some of this...)
