Fieldpoint cFP-1804 cFP-RTD-122 cFP-RTD-124: Who guarantees the sample speed ?

[hemmerling]

hello

 Fieldpoint cFP-1804
with cFP-RTD-122 cFP-RTD-124:

Who guarantees the sample speed ?
So I was told that the RTD modules have an all-channel update period of 1.08s ( = 0.93 Hz ).

So if I use the dumb backplane cFP-1804,
so no CPU on the Fieldpoint system,

must the DAQmx driver  on the Labview computer ask each time for the new samples,
or does the driver get the samples sent automatically ?

And if so, can Labview "wait" in a language construction called "hardware loop" for the receiving of new samples = synchron sampling with 0.93 Hz,
instead of looking all 1 second for new data = asynchron sampling ?

I assume if the fieldpoint system has a CPU onboard, its all different.
Here the fieldpoint system is JUST used for data aquisition ( in a solution advertised for temperature measurement, for a project ).

Sincerely
Rolf

[Wayne.C]

Rolf,

The update rate of 1.08 seconds refers to how fast/often the the A/D conversion is made in the cFP-RTD-124 module itself.  This speed is determined by the hardware and is a published spec from NI.

You don't use the DAQmx driver to read cFP I/O.  You will need to use the "FP Read.vi" to get the information from the module back to the computer.  I usually write a loop that reads the FP I/O at a set interval and make sure the interval is longer than the update rate.

The only difference between what you are doing and having the processor module, is the location of the code that is reading the I/O modules and how fast the overall process runs.

Think of it this way, the I/O module scans the channels and performs an A2D conversion of each input.  The module then makes this new data available via the backplane to a "Read" request from a program running either on a cFP processor module or a program running on another machine.  The overall speed of the final system is a combination of the update rate of the I/O module, speed of network communication with the backplane and speed of the actual program. 

Hope this helps.

[hemmerling]

Thanks alot,
this is exactly what I expected for process control 🙂

Now with 0.93 Hz update frequency for the cFP-RTD-12x,
it is obvious no problem to update the FP-driver on the windows computer.

But what is with a standard analog module cFP-AI-112,
which has a (maximum ?) scan rate of 10 KS/s ?
Where can I set this scan frequency, with the cFP-1804 hardware  ?
Can I set this in the cFP-AI-112,
so that for serial communications, this sample rate can be reduced ?
Or is this set in the FP driver, which itself asks the cFP-AI-112 to send the data ?
With cFP-AI-112, can I set that just x of 8 channels are actively reporting to the Labview computer ?

What is if there are several cFP-AI-112 modules, how can the bottleneck RS-232 be handled ?

How can I calculate how much data samples can be transfered,
so how much RS-232 traffic does the cFP-AI-112 cause ?

Is there much difference if the broadband Ethernet connection is used instead of RS-232 ?
Is there any difference for the Labview  application, if either RS-232 or Ethernet is used ?

I read here
http://digital.ni.com/public.nsf/3efedde4322fef19862567740067f3cc/0d418c9097f14e99862567c300666568?O...
[quote]
The network controllers (FP-160x, (c)FP-2xxx) do not operate in a manner similar to the FP-1000/1001 in that the FP-1000/1001 will respond to poll requests by sending data when they are polled. In order to minimize network use, the network controllers only transmits data when it changes. Basically, the network controllers operates as a data server. The computer will subscribe to input or output items (based upon your IAK file) (Note: with the network controllers it is more efficient to delete unused channels out of the IAK file). When the computer subscribes to an FP-160x's output item, the computer will only send data to the network controllers on change. Likewise, the network controllers will only transmit data from input channels to the computer when on change, so a poll by your program to the network controllers actually goes through Measurement & Automation Explorer (MAX) or FieldPoint Explorer, which gives it the most recent data that it had received and does not physically poll the device. In the case of the FP-1000/1001, it would actually poll the device.
[/quote]
Does this mean that if an analog or temperature value does NOT change, the known value is NOT transfered by the serial RS-232 line, to minimize the traffic over RS-232 ?

Btw, what is your typical scan rate, for the Labview "timed loop" ? 1 Hz  ( so just little more than the 0.93 Hz update frequency ), or much more ?

Sincerely
Rolf

[Wayne.C]

Rolf,

I prefer TCPIP communication via ethernet.  Fast and very reliable.  One way to transfer data a bit faster is to read an entire module with a single read.  This returns an array of data. 

I read data from the cFP only as fast as I need it.  If I were reading thermocouples I might only read the cFP module once every 10 seconds.  If I'm trying to measure 60Hz current, I decide the minimum number of samples I need per line cycle and that determines how often I read the cFP.  Remember, I'm using a regular backplane with the processor module installed and I don't have all the communication delays.  Also remember that the rate at which I read data from the cFP module is totaly independant of the "Update Rate" of the module. 

Following is from the Operators Manual for the AI-112 module.  Note that the filter settings for modules are set in MAX.

Update Rate

The update rate of the [c]FP-AI-112 is determined by the filter

settings and the number of input ranges in use. Higher filter

settings and fewer input ranges yield faster update rates. Use the

following formula to determine the approximate update period:

1

(

number of channels with 50 Hz filter) × 0.061 s +

(

number of input ranges with 50 Hz filter) × 0.12 s +

(

number of channels with 60 Hz filter) × 0.051 s +

[Wayne.C]

Rolf,

My mistake.  Here is complete calculation for "Update Rate"

Update Rate

The update rate of the [c]FP-AI-112 is determined by the filter

settings and the number of input ranges in use. Higher filter

settings and fewer input ranges yield faster update rates. Use the

following formula to determine the approximate update period:

(

number of channels with 50 Hz filter) × 0.061 s +

(

number of input ranges with 50 Hz filter) × 0.12 s +

(

number of channels with 60 Hz filter) × 0.051 s +

(

number of input ranges with 60 Hz filter) × 0.10 s +

(

number of channels with 500 Hz filter) × 0.007 s +

(

number of input ranges with 500 Hz filter) × 0.012 s + 0.16 s

=

Update Period

[hemmerling]

Thanks alot,
especially because your first message and the last 2 caused different conclusions.

First of all, my task is to find out if the analog module can do scans with 10 Hz,
as this is a mandatory project goal for a (quiet unusual ) PT100 temperature measurement.
So the original temperature module just offers 0.93 hz, while the analog module offers up to 10 Khz and might be combined with an extra PT100 signal conditioning unit.

1)
So the internal update rate of the modules is either

For the analog module
10KS/s, while the "effective scan rate"  can be calculated by your formular, and the parameters of this scan rate are SET in the module's configuration parameter.
Hmm...
8 (channels with 50 Hz filter)  * 0.061s  +
1 (input range, as all channels have similar input sources) * 0.12s
+ 0.16
= 0.728 s.
But this is 1.37 Hz

8 (channels with 500 Hz filter)  * 0.007s  +
1 (input range, as all channels have similar input sources) * 0.012s
+ 0.16
= 0.228 s.
But this is 4.38 Hz

Did I calculate right ?

I did not found this in the "operation instructions" manual, which is available online at
http://digital.ni.com/manuals.nsf/websearch/FFFB09E5F6D03B7D86256FB2006A612E

But there I read
***
Filter Settings
The cFP-AI-118 has a 10 Hz digital sync filter that provides 55 dB
of rejection at every multiple of 10 Hz ±0.01%. Each channel can
be configured separately in software for 10 Hz filtering or no
rejection. The filter can greatly reduce 50/60 Hz noise from nearby
AC power lines.

Sample Rate
The all-channel sample rate of the cFP-AI-118 is 10.42 kHz with
no filtering. I fyou enable filtering on any channel, the all-channel
sample rate is 3.45 kHz.
***
Comment #1: From my point of view, a digital 10 Hz filter makes scans with more than 5 Hz useless,
as 10 Hz is the cutoff frequency, and just signal frequencies below 10 Hz can contribute to the fourier sum signal.
Am I right ?

************************************
So where is the power to do scans with 10 KS/s, if I calculate scan frequencies of 1.37 or 4.38 Hz ?
************************************

2)
Btw,
According to your formular,
cFP-AI-118  and cFP-AI-112 have input filters of 50 Hz, 60 Hz, 500 Hz,
Comment #2 : This limits from my point of view, the maximum detectable scan frequency to 1/2 * 50,  1/2 * 60, 1/2 * 500 Hz, as all frequencies "above" are cut off by the filters ?!
Am I right ?

3)
For the temperature module
0.93 Hz effective scan rate for all channels
- which is FIXED  ( and can not be set, if I interpret the data sheet right ) ?
- or it depends on the number of channels used ( with a formular similar for the analog module ) ?

4)
So neither a Fieldpoint processor ( I don´t have such ) nor the Windows FP driver nor the Windows Labview  application knows the parameters or acts according to it.

But most important, the modules don´t send data by the serial device, if not asked. And they even don´t send the data to the Fieldpont processor module ( I don´t have such ).

And then the developer DECIDES by writing the Labview software ( on the Fieldpoint processor module, or on the Windows computer) how often he wants to GET the scanned data.
This Labview scan rate on the Windows computer causes communication by the serial or Ethernet connection, while with a Fieldpoint processor module, this scan rate is less important.

Am I right so far ?

5)
In my case, the vendor of a certain motor testbench JUST supports the serial communication :-(.

So the question still is:
Now if I scan the analog module, by using the FP_read,
how many data items can I transfer with the serial device ( with 115.200 bps =
14400 bytes /s )?
****************************************************************
So I would like to get an answer like this:
Each analog item or temperature item is 16-bit = 2 bytes, plus there is an overhead by the protocol, so that with 115.200 bps, you can transfer the

xxx samples  per second
and NOT MORE.

And if I use 4 / 8 channels with 1 module, or 10 channels with 2 modules, I must divide this  xxx by the number of channels.

The question is, what is the number xxx?
****************************************************************
Of course with Ethernet, the barrier is much higher and probably above the 10KS/s level even with many modules...

Of course Labview scan  rates for the temperature higher above the INTERNAL scan rate of 0.93 Hz are useless,
and of course I can transfer the data of 4 or 10 channels, with 1 Hz scan rate, by the serial device.
But with the analog module, with many channels, not the 10KS/s but the serial device is the bottleneck, I assume...

Sincerely
Rolf

Message Edited by hemmerling on 01-20-2007 02:09 AM

[hemmerling]

How to get a high speed sampling ?

According to the formular

8 (channels with 50 Hz filter)  * 0.061s  +
1 (input range, as all channels have similar input sources) * 0.12s
+ 0.16
= 0.728 s.
But this is 1.37 Hz

8 (channels with 500 Hz filter)  * 0.007s  +
1 (input range, as all channels have similar input sources) * 0.012s
+ 0.16
= 0.228 s.
But this is 4.38 Hz

Did I calculate right ?

I did not found this in the "operation instructions" manual, which is available online at
http://digital.ni.com/manuals.nsf/websearch/FFFB09E5F6D03B7D86256FB2006A612E

But there I read
***
Filter Settings
The cFP-AI-118 has a 10 Hz digital sync filter that provides 55 dB
of rejection at every multiple of 10 Hz ±0.01%. Each channel can
be configured separately in software for 10 Hz filtering or no
rejection. The filter can greatly reduce 50/60 Hz noise from nearby
AC power lines.

Sample Rate
The all-channel sample rate of the cFP-AI-118 is 10.42 kHz with
no filtering. I fyou enable filtering on any channel, the all-channel
sample rate is 3.45 kHz.

****************

So  with filters activated, there is not much more than 1,37 and 4,38 Hz ?

[Wayne.C]

You are forgetting that the cFP is described by NI as a Programmable Automation Controller (PAC).  It's strengths are that its hardware is designed for an industrial environment and that it runs RealTime.  It's I/O modules are geared more towards process control and not high speed data acquisition.  If the calculations for the AI-118 module update rate indicate that this speed is not fast enough for your application,  you will have to consider some other DAQ hardware.

[hemmerling]

Thanks alot,

but the ONLY reason to consider the FP is that it is an offer of a motor testbed manufacturer,
and we must decide to take it or leave it, by a reason.

Nevertheless, if there is no typing mistake,
with ACTIVATED filters its not usuable for the UNUSUAL 10 Hz scan frequency demand.
So the ONLY solution is to DEACTIVATE the filters.

So I might switch off the filters and especially the 10Hz software filter,
and then I get 10 KS/s, right ?

its just that then I get all the peaks and spots caused by electric magnetic noise,
and thats what the filters are for to avoid it.

We were told that there is no (much) of that EM noise... but who knows. I won´t lay my hand in fire that the temperature signal is then VERY noisy, maybe...
just to get the 10 Hz scan frequency.

Sincerely
Rolf