Re: [foxboro] Analog values from Triconex to I/A via Modbus Int 30
- From: "Gregory A Hurwitt" <hurwitg@xxxxxxxxxxxxx>
- To: foxboro@xxxxxxxxxxxxx
- Date: Wed, 9 Feb 2005 14:23:59 -0600
I have looked at some previous configuration work and found the answer to
my own question:
1. Set MinSpan/MaxSpan in Triconex configuration to the desired
engineering unit range (i.e., set to match LSCO1/HSCO1 in the I/A AIN
block). This tells Triconex to scale the value before transmission over
Modbus such that 0-32767 counts corresponds to the desired engineering
range.
2. Use SCI = 12 (0 to 32767 counts) in the AIN block.
This method allows communication of real numbers (to a precision of 1 part
in 32767).
Thanks to those who replied.
Greg Hurwitt
|---------+------------------------+---------------------------->
| | |"Gregory A Hurwitt" |
| | |<hurwitg@xxxxxxxxxxxxx> |
| | |Sent by: |
| | |foxboro-bounce@xxxxxxxxxxxxx|
| | | |
| | | |
| | |02/09/05 11:32 AM |
| | |Please respond to foxboro |
| | | |
|---------+------------------------+---------------------------->
>---------------------------------------------------------------------------------------------------------------|
| To: foxboro
|
| cc:
|
| Subject: [foxboro] Analog values from Triconex to I/A via Modbus Int 30
|
>---------------------------------------------------------------------------------------------------------------|
|---------+------------------------+---------------------------->
| | | |
|---------+------------------------+---------------------------->
>---------------------------------------------------------------------------------------------------------------|
|
|
>---------------------------------------------------------------------------------------------------------------|
Hello list,
I am trying to communicate a few analog values from a Triconex system
(Tricon V9.x) to I/A (V6.x) via a Modbus Integrator 30.
These values are calculated in the Triconex, and stored in read-aliased
real memory locations. (Alias numbers beginning at 33001.)
My MDSCAN block is set up with DATYPE = 4 (read-only analog data beginning
at address 30000). I am aware that only the integer portion of my value
will be communicated, and this is perfectly acceptable for my application.
The problem I am having is that a value of 0.0 at the Triconex is
interpreted as 819 raw counts in my Mod Int 30 AIN block. This happens
regardless of the setting for SCI in the AIN block. (Makes sense; signal
conditioning is applied *to* raw counts to generate the measurement value.)
A value of 32767.0 at Triconex is seen as 32767 counts at the AIN block.
The "offset" in the raw counts is linear between these two endpoints, so I
have been able to come up with KSCALE/BSCALE factors that allow me to work
around the problem. IMO, this is a very ugly workaround.
Of course, anyone who has worked with Triconex at all knows that 0-20 mA
analog inputs are interpreted in the system as 0-4095 counts, and that 819
counts is the exact value for 4 mA. It seems like I/A is trying to "help"
me by accounting for a 4mA zero signal, although it does seem backwards
(819 in Triconex should be interpreted as 0 in I/A, not vice versa as is
happening). However, none of this applies to my situation, since I'm not
looking at an analog input in counts in Triconex, but rather a calculated
value already in engineering units. I simply want to communicate the
existing value with no modification.
Is there an SCI setting that "reverses" the 819 count offset? I sure don't
see one in the documentation. Or do I have something configured
incorrectly? Or is this a known limitation of the Modbus Integrator 30?
_______________________________________________________________________
This mailing list is neither sponsored nor endorsed by Invensys Process
Systems (formerly The Foxboro Company). Use the info you obtain here at
your own risks. Read http://www.thecassandraproject.org/disclaimer.html
foxboro mailing list: //www.freelists.org/list/foxboro
to subscribe: mailto:foxboro-request@xxxxxxxxxxxxx?subject=join
to unsubscribe: mailto:foxboro-request@xxxxxxxxxxxxx?subject=leave
Other related posts:
