Setting Waveform Chart History Length to very large value crashes Labview

[barrysilver1]

Hi ...

The attached .vi is a simple read from PCI and log into TDMS file. 

I set the chart history buffer to some very large number (2e6 or thereabouts) and now the .vi refuses to open without crashing the entire Labview environment.

Would someone be able to open it and reset the chart history buffer to some low number so I can continue please ?

I should mention it was compiled in Labview 2021 SP1

[dkfire]

Try clear your LabVIEW cache, under Tools->Advanced->Clear Compiled Obejct Cache.. 

[barrysilver1]

I cleared the cache as suggested and tried to load the .vi into Labview ... it still crashed !

[GerdW]

Hi barry,

---

@barrysilver1 wrote:

I set the chart history buffer to some very large number (2e6 or thereabouts) and now the .vi refuses to open without crashing the entire Labview environment.

---

Why did you do that?

---

@barrysilver1 wrote:

Would someone be able to open it and reset the chart history buffer to some low number so I can continue please ?

---

Nope, crashes for me, too.

But why don't you pick an older version of this VI from your SCC tool? (Or from an backup?) That is the most simplest solution to your problem!

When you don't have a backup: recreate the VI from scratch…

[Bob_Schor]

Ah, "We apologize for the Inconvenience" crash.  It "blew up" for me with LabVIEW 2021.  Your best bet is to reload from backup (you are using some form of Version Control, I hope ...).  Otherwise, 

1. Reconstruct the VI from memory.
2. Don't use more than a few MB for Chart History.  It will keep growing, and growing, and growing ...
3. What is the real problem you are trying to solve?  There may be other ways to do it (including streaming to disk and keeping in view "the last XXX Hours/Minutes/Seconds" of data on the screen).

Bob Schor

[barrysilver1]

Hi All

I was originally trying to extend the waveform plot time (to an infinitely continuous stream )  as it seemed to be stopping after a while (which was sample rate dependent).

I originally had the buffer at 30000 and so was just seeing how much it would take (before it blew up!).

Damn ! ..... would somebody be able to debug the .exe that I built from this .vi (.exe attached) and get a copy of the block diagram ??

Thanks ...

[Kevin_Price]

No solution here, just some info for the future.

When a chart takes a waveform as a datatype, its history length refers to the # of distinct waveforms it should retain, not the total # of samples.  Each of those waveforms may contain 100's of samples in an array of 8-byte DBLS, so the memory space required quickly becomes 1000x or more the number used as history length.

-Kevin P

[barrysilver1 · Accepted solution]

Lesson learnt ! .... Thanks for all your useful replies !!