FieldPoint Family

cancel
Showing results for 
Search instead for 
Did you mean: 

Fielpoint vs cRIO

Does anyone have any experience of working with both cRIO and fieldpoint? If so, do you have any comment on the text below? A sales guy is trying to sell us cRIO rather than fieldpoint for a system that we are developing. I am trying to get some further info before we proceed. By reading the datasheets for the fielpoint I/O modules, it would appear fieldpoint would cope with our requirements, however the sales rep we have dealt with (non NI) has suggested otherwise.
 
Any comments appreciated.
 
 
 
 
 
Compact RIO vs Fieldpoint.

Comparing Compact RIO with Compact Fieldpoint - The important parameter is the sample rate.  Compact RIO is the correct choice for the sample rates that you specified (50 to 100 samples/s).

Compact Fieldpoint is designed primarily to acquire static data (data that is changing at rates of a few Hz or less).  The overall sample rate that can be achieved in a Compact Fieldpoint system is a combination of the internal update rate of the Module itself, and the rate that data can be transferred from all the Modules in the bank to the Controller.  This means that the maximum overall sample rate is almost always less than the maximum Module update rate, and also that it can vary, depending upon the number of Modules and channels being sampled, and the load on the Controller.

This mechanism in Compact Fieldpoint also means that the point-to-point timestamp is typically subject to significant jitter once the overall sample rate exceeds about 20 samples/s.  This is a serious issue if you intend doing any sort of time series analysis on the data.  One client in particular  ran into just this problem - they used Compact Fieldpoint to acquire water wave data at sample rates of around 20 samples/s, and performed frequency analysis looking for events in the range 2-3 Hz.  The imprecision of the timestamp, even at only 20 samples/s, was enough to introduce false frequency artifacts into the sampled data.

For these reasons Compact Fieldpoint
is not recommended for data logging applications above about 20-30 samples/s, particularly where the data may be subject to time series analysis.  There are Compact Fieldpoint analog acquisition Modules with internal update rates of 100 samples/s or more, but we would only use them for single point PID control, NOT for time dependent data logging.

Compact RIO on the other hand, is designed to acquire data at much higher rates than Compact Fieldpoint (up to 100 ksamples/s), and the sample rate is not affected by the number of channels being sampled, or the Controller load.  The precision of the point to point timestamp is maintained at all sample rates, so the data is immune to time series analysis artifacts due to sampling jitter.

The sample rates that you specified (50 to 100 samples/s) are at best marginal for Compact Fieldpoint, and at worst could not be guaranteed under all sampling conditions.  Even if these rates were consistently achievable using Compact Fieldpoint, the timing jitter would be significant, making the data unsuited to time series analysis.  With Compact RIO we can guarantee the required sample rate under all conditions, and the data will be amenable to time series analysis.
0 Kudos
Message 1 of 4
(7,385 Views)
It would be helpfull to have to same description of your system as the sales rep you asked.

If you are sampling the data to be stored for analysis, the points he addresses look at least somethong to think about. Do they apply to my situation?
If you are doing control and some jitter doesn't affect the correct working of your system, maybe Fieldpoint is enough.

Developing an application for Fieldpoint is a two step process (One app on the PC and one in the Fieldpoint) and for the cRIO three step (one in the FPGA, one in the Controller and one in the PC). The PC app is optional if the system runs stand-alone.

It comes down to the fact that sample rate is not the only parameter that decides the best solution.
Regards,
André (CLA, CLED)
0 Kudos
Message 2 of 4
(7,367 Views)

Andre,

Thanks for the response. Please find a brief description of our system below.

We have sensors in the field wired back to a local junction box with either FP or cRIO controller. We are logging the values from these sensors continuously @ 40 samples per second and storing the values on the controllers buffer. When a certain event occurs, ie a valve moves, we want to download all the stored values for the timestamps of the event from the controller to a server located back in the control room which will enable us to perform our analysis. We get the timestamps of these events from the plant control system- this is how we know what data to download from the controller.

0 Kudos
Message 3 of 4
(7,365 Views)
Since you don't perform analysis on the 40 sample/s data, but only a subset that is related to an event. The argument of the sales rep about analysis errors is not important to my opinion. The actual data logging rates in the control system are much lower that 40 Hz.
Regards,
André (CLA, CLED)
0 Kudos
Message 4 of 4
(7,363 Views)