FP Read time stamps are skipping

[ChrisFarmerWIS]

Hi!

Using:
LabVIEW RT 7
FP2010

I am experiencing a problem with the timing of the FP Read VIs. My application is calling the FP Read.vi in a 5ms time-critical while loop. Each sample of data is being placed on an RT FIFO. In another VI, the data is being read off the RT FIFO, and being sent to the host PC application for processing.

I have noticed that every now and then, the time stamp seems to skip 1 second!

Now when I saw this problem originally, I assumed that it was to do with the time server. At which point, I ceased to use the time server, and manually set the Date/Time instead (before the start of a test).

This seems to have improved the time stamps, however I am still witnessing this 1 second jump every so of
ten.

From my calculations, I have more than enough RT FIFO buffer for the data required, so I don't believe I am overwriting the RT FIFOs.

Has anyone seen this problem before?

Thanks
Chris Farmer

[ketch]

Hello my name is Cacciacarne Stefano and I have the same problem of you.
I have a critical READ.vi (40 ms) which send by RT FIFO the data to a non critical vi (150ms).
This one send by Data socket the data to an host vi. by TCP/IP.
Some time I have 1 sec of shift on time tick time of Read.vi timestamp.
Have you received any answer by NI?
Can you help me?
Sorry for my English

[Ben]

Hi Chris!

You wrote "the time stamp SEEMS (emphesis added) to skip 1 second!"

I would suggest you first determine if you are getting bad times or if the TC loop is being stalled/interupted.

1) Track the "get time in seconds" along with the FP time stamps and see if they show the same issue.

2) Drive your FP input with an external signal (like a sine wave) and log that along with the time stamps. The acquired data will confirm if the issue is with the determinism of the TC loop or associated with the time stamp value.

Post a 6.1 version and I will be happy to look at the code.

Trying to repay your service,

Ben

(Ben Rayner, Data Science Automation)

[ChrisFarmerWIS]

Hi Cacciacarne,

Good to know that somebody else is experiencing a similar problem!

I have spoken to NI Australia, and they are investigating the issue for me.

Regards
CF

[ChrisFarmerWIS]

Hi Ben,

Thanks for your prompt response.

Firstly I should point out that I am not Chris Farmer who works for NI. I am another Chris Farmer, who works for an alliance member in Australia! Sorry for any confusion.

I will try some of your suggestions. Could you please tell me how the TC loop could be stalled/interrupted? I thought that the TC loop pre-empted any other process running on the FP module?

Thanks again.

Chris Farmer
Electronic Systems Engineer
CPE Systems Pty Ltd

[Ben]

Well that clears up why he would be posting here (aside from doing a survey).

"Could you please tell me how the TC loop could be stalled/interrupted?"
The short answer is memory management (there are others).

I can not list all of the possiblities in a short answer but if you post your code (in LV 6.1) I will take a quick look and let you know if there could be other issues.

Ben

[Ben]

Please post a LV 6.1 version and I will take a look.

Ben

[Ben]

Did you figure this out?

Ben

[ChrisFarmerWIS]

Hi Ben,

We replaced the timestamp with a "Get Date/Time in seconds.vi", as you suggested. The problem still occurring, without extensively testing it.

I just created a 6.1 version of the code for you, but this web site is not working very well, and won't let me attach the files!!!

I'll try in a couple of days, and let NI know that it isn't working at present.

Thanks
CF

[Ben]

I will watch for your post!

I do alot of LV-RT work and I do my best to keep up on the details.

Ben