Does LAbview 6 ever have problems dealing with serial inputs sharing the same interrupts?

[lobsterbubble]

I'm running Labview6 in win95. My pc104 computer has 8 COM ports and a DAQCard-516. I'm acquiring data from 4 serial conductivity cells which stream data continuously at 32Hz, 1 serial GPS streaming continuously at 1Hz, 1 serial tilt sensor which is sampled at 5Hz, and 1 pressure sensor sampled at 5-10Hz through the DAQcard. The problem I'm running into is that Labview seems to not be able to keep up with the data acquisition of all the instruments. Labview will be ok for a few seconds and then start corrupting the data. From diagnostics I've tried, the problems seem to deal with shared interrupts among the COM ports and Labview not being able to process the data fast enough. I've
attached my code and a data file from one of the conductivity cells. You will see that the data looks ok for a whil and then quickly turns to garbage. The following is a list of the COM ports and their respective interrupt #'s and instruments.
COM1 int4 tilt sensor
COM2 int3 conductivity sensor1
COM3 int4 conductivity sensor2
COM4 int5 GPS
COM5 int5 conductivity sensor3
COM6 int7 conductivity sensor4
COM7 int5 nothing yet. will mirror the output of COM4 (GPS string) to be coded.
COM8 disabled
DAQcard int9

This is the only configuration I've gotten all the instruments to be recognized by Labview and Windows. I've connected the conductivity cells such that none of them share the same interrupt, but 2 share interrupts with slower data rate instruments.
Any help or thoughts would be great.

I tried attaching the files, but the website wouldn't let me. Please send me an e-mail address I can send the files to.

[Rob Cole]

If you answer your own question, you will be given the option of attaching a file with your answer.

Something to check: Have you disabled the printer port on your PC? WIN95 periodically polls the parallel port to see if a printer has been connected (int7 usually).

Does the system work with a subset of the sensors and com ports? Simplify the system (no shared interrupts) and see if the problem persists. Look to see if it's a shared interrupt problem or something else.

I'll be looking for the data file to see what it looks like.

Rob

[lobsterbubble]

I did get to the page where it asks for the files I want to attach, but I get an error after I try to attach the files. I tried again today with the same error message.

I have disabled the printer port (int7) in the BIOS. That was the only way I could get all conductivity cells on different interrupts. The data seems to geet corrupted more often if conductivity cells are sharing the same interrupt. The problem with configuring the system so that none of the instruments share an interrupt is that the extra COM ports I've mounted on our computer will only work if COM5 and 7 share interrupt 5 and COM6 and 8 share interrupt 7. They will work on int9, but the DAQ card has that spot. They don't work on int10 or higher. Having 6 out of 7 instruments with s
erial interface prevents me from doing that.

From all the different configurations I've tried, I'm positive it's 2 problems. Sharing interrupts causes one and Labview not being able to keep up with all the data is the other. The code started off small with only a couple of instruments. Code was then added to accomodate more sensors. I didn't have problems when I only had a couple of instruments, but when I got to the point where the VI was acquiring just 4 conductivity cells, the data would go bad even when I had all of them on different interrupts.

I will try again to attach the files on a different computer. If it doesn't work, i'll wait to hear from you again to see how to continue.

[Rob Cole]

I've had a little look at the data and the VI (though it won't work on my system). I think that you're right about LabVIEW not being able to keep up. Especially when pitch and roll happen.

As a check on this - Have you tried breaking the links to write data to the front panel or to write to the files? Just as a check on whether it can keep up either way. Both of these operations take a certain amount of time. You may just have to look for more efficient methods of doing what you're doing. Or showing a subset of the data (instead of everything on screen). I know - a bunch of them are inputs that are just looked at once - those should be fine.

Other than that, I must admit that I don't know. Shared interrupts have always been a problem in t
he Windoze environment. I have one 8-port serial card here, but the card uses one interrupt with different IO ports for each serial port. Works well, but I haven't tried to run all ports at higher speeds simultaneously.

I hope that you find the problem and keep me informed as to what you find.

Rob

[lobsterbubble]

It worked!