cFP-2020 Locks up after 1-3 hours

[Ben]

Please share the results.

Ben

[seanpatrick]

OK it has been a while since I gave an update. As far as we can tell the cFP looses contact with the host after a period of time.. About half of the time when it looses contact I get a 4 flash code on the controller but no one seems to know what that means. The local NI rep brought his 2020 in and it did the same thing. So I guess that proves that it is not hardware. This means that I am going to have to re write the code. I am going to try vi server but this is kind of leaving me a bit upset. Especially since the existing program was set up great except that the cFP keeps dropping dead. Oh and for the record we have checked the memory a couple of ways and it always stays around 14.5 Mb free.

[Ben]

The utility I shared with you also should the largest contiguous block.

How does that look?

It is possible to have free memory but not have a contiguous block large enough to complete an operation.

What happens if you replace your RT loop VI with a dummy?

Ben

[cous]

Hi All,
This post hasn't been touched for a while, so I'm updating what happened.
We determined that the issue was mostly code, not hardware related, so we reduced to code to use VISA and remove some of the queues that were used. Those changes resulted in a working version of code that didn't crash. I'm not sure how Sean eventually resolved the issue, but it looks like isolating a portion of problematic code helped resolve the crashing problem.
Marcus G.
Applications Engineer
National Instruments

[cs42]

Hi all,

unfortunately I just stumbled over this thread -- a little late. But my first thought after reading the first messages was: memory fragmentation.

Actually in my experience the free memory isn't very often a problem, but the fragmentation of the memory over the time. So the provided tools might tell you a constant amount of free memory but if you're not cautious in avoiding using dynamically sized arrays, strings and queues other than the RT ones you'll end up with unusable free memory because of fragmentation. One sign of fragmentation in my experience is, that such a program runs well for quite a while and a very similar progam (if possible) will run on a PC for ever without any problem...

So what to do? Read the RT manuals; don't use dynamically resized array; if an RT specific function is available use this and not the "normal" counterpart; don't use the standard "merge errors.vi", because this does a dynamical array allocation (write your own; the LV 7.1 version is better)...

Just my 2 cents,
Carsten