NI 9481 becomes unresponsive overnight

[NotanNIguru]

This is to control an environmental chamber. Feature I am having trouble with will allow for the user to set a time to have the chamber turn on in the morning, hence saving power. My problem is the program works fine, until the computer sits over night. When I come in the next day the temperature on the Watlow F4 has changed (this means the VI ran, should eliminate sleep/hibernate issues) but the relay did not close. This is running on a PC with Windows XP. I have posted the project. Comms with the Watlow are through a USB to Serial adapted, the 9481 is connected to a cDAQ 9171.

[Yupeng_Jia · Accepted solution]

Hi NotanNIguru, it seems like a memory leak in your program. What’s the memory size of your computer and what’s the OS?

There are a few changes you could make with your code:

1. Check if you have closed references properly. Opening references repeatedly without closing them may cause a memory problem because each reference requires memory.
2. There are a lot of coercion dots on your block diagrams. When you see a red dot appear by the input of a function, this means that the function is expecting one type of data, but receives another. LabVIEW converts these for you automatically, but it does have to create a buffer in memory to change the data type of the value. You can remedy this issue by right-clicking on the control or indicator on your front panel and selecting Representation and changing the representation to match the other data type going into the function.

There are a couple of articles talking how to improve LabVIEW performance:

Improving the Performance of Your LabVIEW Applications

Tips and Tricks to Improve LV Performance

[AnalogKid2DigitalMan]

Are you using the 9481 to apply power to the chamber?

If so, what is the voltage and current requirements of the chamber?

-AK2DM

[NotanNIguru]

I am not dirrectly powering the chamber, I am merely replicating the actions of the physical buttons to allow remote opperation.

I am passing 120VAC to a relay, the closure only needs to be made for a moment.

[NotanNIguru]

This is running on Windows XP with 2G of Ram.

I did have a memory waterfall, I found after I posted so here is the update:

1) Good call. I was not closing the DAQmx properly, fixing that seems to have controlled the leak mostly...

2) Good Idea, I will look into this. There is still an increased memory allocation of about 10K every minute or so, the coersion might be the culprit.

I ran the program over the weekend (After DAQmx being properly closed) and still no trigger. I will update the representation of th inputs into functions and try again tonight.

Thank you for all your help.