FieldPoint Family

cancel
Showing results for 
Search instead for 
Did you mean: 

Ethernet communication problem after upgrading to cfp-2220

I will try the console as I have never used it.  I can say for sure that there is a major issue related to timed loops in the cFP 2220 OS.  This is based on my cFP 2120 code never showing this issue and when ported to the cFP 2220 the cFP and the Touch Screen would loose TCP communications within minutes.  After replacing timed loop that contained a TCP function (TCP Read) the problem still existed.  After replacing all timed loops with While loops and msec delays The problem has disappeared.  These other loops were logging, UDP messaging, main loop and blinking status LED.

 

So let's say that it is a burst of high CPU load that is causing the issue.  Why would the OS disable TCP communications and require a reboot to get it back?  Also Why would UDP continue to work?  

0 Kudos
Message 11 of 13
(2,972 Views)

One other thing came to mind regarding your application.

 

If you are constantly opening and closing TCP connections in quick succession you may be running into a bug in LV RT.

 

Take a look at this post from earlier this year.

Greg Sussman
Sr Business Manager A/D/G BU
0 Kudos
Message 12 of 13
(2,970 Views)

I have had many go-rounds with VX Works and how it handles the TCP communications. As to why communicaions do not come back when the TCP thread gets re-enabled, I can't answer, however I have seen that behavior myself on a couple of systems I did about 2 years ago. The best way I could describe it was when the TCP thread was suspended something happened to connection between VX Works and LabVIEW RT.

 

As far as my client computer was concerned, the TCP socket was still open. The client could send data to the RT controller and not receive any sending errors.

On the LVRT side, the TCP read functions never returned an error, but also never returned any more data after the TCP thread was resumed. It was as though the link between VX Works and LVRT was partially broken as no further data was passed, but the conections still indicated open and good.

 

Due to the design of the RT program I was not able to test if the RT side had the ability to send data out in this condition.

 

My solution, although it was not much in the way of fixing the problem, was to modify other areas of my code to preclude high CPU demand and thus prevent the TCP thread from being suspended. As long as the thread remained active, my application worked fine.

 

I worked with one AE and even the PSE at trying to duplicate the behavior, but was never successful. It seemed sporadic at best, sometimes the connection would restore fine, other times it would come back in this partially broken state.

Greg Sussman
Sr Business Manager A/D/G BU
0 Kudos
Message 13 of 13
(2,966 Views)