FieldPoint Family

Showing results for 
Search instead for 
Did you mean: 

cFP Processor forgets parameters across power cycle

having important problem right near deployment. Any insights helpful:

Two separate cFP processors (cFP-2110 & cFP-2120) are forgetting device range assignments set by MAX across DC power cycles. We originally thought it might be DC power quality problems, but now every lead into or out of the cFP processors is ferrerite choked, the DC is regulated and bypassed by 8KuF near input terminals.  Strangely a third cFP-2120 does not seem to be forgetting its parameters, but I don't want to deploy that one because of an earlier problem that makes me distrust it.

The parameters, specifically, are for the range (+/- 10.5 volts) on several of the installed cFP-AI-118s. The range in the cFP memory changes from our desired +/- 10.5 V to the default +/- 1.05 volts.

Has anyone else seen problems with loss of parameters apprently set by MAX when the cFP processor is power cycled?


Jerry Whelan

0 Kudos
Message 1 of 4
Additional information: I took all involved cFP processors into begnin laboratory environment, different backplane and set of modules. In that environment, all processors remember their programming across power cycles. Looks like there is something in the application environment that is causing the parameter loss. Suspect might be backplane metal box is not at the same ground potential as rest of system.

Is there any knowlege of, or experience with, susceptability to malfunction under the circumstances of the cFP backplane metal box is electrically "noisy" with respect to signal / power grounds of the processor ?
0 Kudos
Message 2 of 4
Hi, Jerry.

It sounds like you need to take a snapshot of the configuration you'd like. This KnowledgeBase talks about the issue and how you can do that.

If for some reason that doesn't help, let me know. Have a nice evening!
Sarah K.
Search PME
National Instruments
0 Kudos
Message 3 of 4
Successful problem resolution, thanks to all who commented. I will briefly list the actions we took to eliminate the reproducible problem, as well as some useful information about programmatically accomplishing what MAX was being used for:
The channel range parameters were stored successfully (after being configured using MAX) into the cFP-2120's flash memory.  The cFP processor would remember its parameters in any other benign environment, but some transient voltage difference that existed in our system between the processor DC power lead pair and the backplane metal box was causing the cFP system to regularly forget random channel gains, setting them back to the first gain in the gain list. On paper, it looked like the backplane metal box was directly connected to the power system's single point ground where the DC power "-" terminal also originated. In practice, some disruptive energy was getting coupled into the several foot long effective path between the processor power terminals and the backplane metal box ground terminal and this transient ocurring at power-off was causing the cFP processor to randomly modify the configured channel gain. No other wierd behavior was observed - at each power-up a fairly complex set of programs successfully downloaded and executed, the system was just forgetting its gains.
That problem is now solved by wire path re-routing, but along the way I discovered, with assistance from NI, that it is possible for a user program to programmatically set the gains of the AI-118, as well as programmatically read-back those gain settings to verify their correctness. For setting gains, the Advanced cFP Palette contains an Ethernet Configure Range VI. This VI does not know about the cFP-AI-118 (too new) but the VI will work successfully IF you are careful to choose "Unspecified Board Type" and insure that you ask for only one of the valid settings for the cFP-AI-118 (as shown available in MAX).
There does not seem to be a provided VI for reading back the configured range of an cFP-AI-118, but here NI tech support was very helpful in providing information about an UNSUPPORTED way of reading back the configured gain: Use a datasocket read to the cFP ethernet address of the form: "fieldpoint:\\\FP\1AI\00\ConfiguredRange"  where 1AI represents module position 1, analog input.  00 represents chanel 0 of 0 through 7. The return is a variant which I converted to a Uint. To interpret the meaning of this integer return, create an indicator from the "RANGE" control that can be found in the NI provided Ethernet Configure Range VI from the Advanced cFP Palette.
This works like a charm: we can read-back present configured gains and set configured gains both programmatically.  Gotta love this stuff ! Thanks again to who, acting in the capacity of tech support, came through quickly when we were in a pinch!
0 Kudos
Message 4 of 4