Re: [foxboro] Multiple MDSCANs in a compound???

Hitesh,

If it is at all possible to concentrate your data on the PLC side you should do
so.  It's normal to have your data sources spread widely over a given address
range.  The common solution is to add a bit of logic at the end of your PLC to
read the values you need to pass to the DCS from the many different locations
where they are used in the PLC and write them to a small block of addresses you
can easily pass to the DCS.  You should do this because Modbus serial comms has
very limited bandwidth.

> Also I have to keep the AINs and COUTs in the same compound. 
> For which the DATYPE would be different than CIN blocks.
So, keep them in the same compound.  Just run your MDSCAN in a different
compound.  For simplicity, consider using a scanner compounds like:

SCAN1
 MDSCAN
 MCIN
 MCIN
 MCIN
 MCIN

SCAN2
 MDSCAN
 MAIN
 MAIN
 MAIN

SCAN3
 MDSCAN
 MCOUT
 MCOUT

PROC1
 AIN
 AIN
 CIN
 COUT

The blocks in PROC1 are your normal loop blocks.  Configure them as IOMOPT = 0,
and point them to the bit in the MCIN or the value in the MAIN.  Give your MCIN,
MCOUT and MAIN blocks address descriptive names and descriptions, since all they
are is the DCS representation of a range of PLC addresses.


Some general guidelines (my guidelines, not necessarily Foxboro rules -- your
mileage may vary):
* Yes, only one MDSCAN per compound.
* Keep your serial wiring between the integrator and the PLC or bridge short so
you can run the link at 19200 baud.
* For each compound with a MDSCAN block, configure the compound and all the
blocks in that compound to run at a single period/phase.  Don't leave the
compound at PERIOD 1 if all the blocks in it are running at PERIOD = 3.
* Consider using a 2 second period (PERIOD = 3) and run your integrator at a .5
second BPC -- This gives you 4 phases to work with.
* Don't run more than 2 MDSCANs per phase.
* Although you can pass a maximum of 128 analog or 2048 discrete values in a
MDSCAN, don't.  If you stick with 64 and 1024 as a working maximum, you will
keep each MDSCAN to a single read or write.
* Writes take more time to process than reads.
* Monitor the IO load and phase loading in the integrator from the Station block
detail display.  Keep the IO load under 75% and watch for phases that are
significantly more loaded than others.



Quoting "Hitesh B. Kakadiya" <hiteshk70@xxxxxxxxx>:

> Hi list,
> I am presently configuring an integrator for communication with GE PLC.
> Unfortunately, the modbus allocation on PLC side is very random and can
> not
> be changed. For example, There are 14 CIN signals in one compounds
> which
> are to be read over the address range 07205 to 09304. This clearly
> exceeds
> PCINPS range of 2048 recommended for MDSCAN block. Also I have to keep
> the
> AINs and COUTs in the same compound. For which the DATYPE would be
> different
> than CIN blocks.
> 
> I am wondering is it possible to put more than one MDSCAN blocks in a
> compound. FOXDOC mentions that MDSCAN should be the first block but
> does
> not specify whether it is the only block.
> 
> Sincerely appreciate your help.
> 
> Thanking in advance
> 
> Hitesh Kakadiya
> I&C Specialist, Colt Engineering Canada
> 
> 
>  
>  
> ______________________________________________________________________
> _
> 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: http://www.freelists.org/list/foxboro
> to subscribe: mailto:foxboro-request@xxxxxxxxxxxxx?subject=join
> to unsubscribe: mailto:foxboro-request@xxxxxxxxxxxxx?subject=leave
>  
>  

 
 
_______________________________________________________________________
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:             http://www.freelists.org/list/foxboro
to subscribe:         mailto:foxboro-request@xxxxxxxxxxxxx?subject=join
to unsubscribe:      mailto:foxboro-request@xxxxxxxxxxxxx?subject=leave

Other related posts: