Skip to Content

SiePortal

The integrated platform for your product selection, buying and support workflow - bringing together Industry Mall and Online Support.

SiePortalDeutsch

Search

Home
Support Support
{{(apiData.Support | filter:{'rank':1}:true)[0].title}} Support
- Menu temporary not available
- Menu temporary not available
- Menue temporary not available
- {{item.title}}

Products & Services
- Menu temporary not available
- Menue temporary not available
- {{item.title}}
Support Support
{{(apiData.Support | filter:{'rank':1}:true)[0].title}} Support
- Menu temporary not available
- Menue temporary not available
- {{item.title}}
MySiePortal MySiePortal
{{(apiData.MySiePortal | filter:{'rank':1}:true)[0].title}} MySiePortal
- Menu temporary not available
- Menue temporary not available
- {{item.title}}
Cart ({{apiData.CartCount}})

Navigation

Open Modbus TCP/IP speed

Created by: ToreK at: 1/21/2011 1:22 PM (7 Replies)

Rating (0)

Thanks 0

8 Entries

1/21/2011 1:22 PM	Rate (0)
ToreK Posts: 13 Rating: (0)	Hello, I am setting up a system using the Open Modbus. Everything is working at the moment, but the system is too slow. My PLC is communicating with an ETU(PC) using a CP443-1. The PLC is the Client. The ETU only have one Port available (502) and only one IP, but there are 6 different slave IDs(Units). All the different slave IDs have the same modbus register setup. At the moment the system is running cyclic. One ModbusCP for each slaveID. Each modbusCP is setup with different DBs (This means that each slaveID gets 3 DBs, One for holding, one for inputs and one for coils. Input register=NA). Each instance of the ModbusCP must be called 6 times to read/write all the requirred registers. This takes in total about 4 seconds in a 20ms OB. 1. Is it possible to do this an faster? 2. Is it possible to control which DB or where in the DB receives data from ModbusCP when UNIT is changed? Same modbus registers from several different slaves should be written to either different places in the same DB or to different DBs by using only one ModbusCP. -Tore-
ToreK Posts: 13 Rating: (0)
Suggestion To thank Quote Answer

1/21/2011 1:41 PM	Rate (0)
Wüstenräuber Gold Expert Joined: 10/3/2008 Last visit: 9/8/2025 Posts: 996 Rating: (129)	Hi, which version do you use? The current version V4.1 is very fast... ETU has 1 ip address and 1 port number. So you need only 1 TCP connection in NetPro and 1 instance of MODBUSCP. During runtime you can set the parameter UNIT to the different 6 values. After each telegram you can BLKMOV the data from slaveID 1 to a defined DB. The other values from the other slaveIDs you can copy to the same DB (to another address in the DB) or to another DB. So the DB, which you parameterized in OB100, you can use as "receive buffer". Wüstenräuber

Suggestion To thank Quote Answer

1/21/2011 2:04 PM	Rate (0)
ToreK Posts: 13 Rating: (0)	I just received the software a couple of weeks ago. In the package there was a previous version and a current version. I am using the one from the perious version which is modbusCP version 1.1 (Version (Header)). This is because i do not think that the CP443-1 is working with AG_CNTRL. The latest verison i can find in the software is version 2.0 of ModbusCP. The package it was delivered in says S7-OpenModBus/TCP, CP V4.x. Where is the version 4.1?
ToreK Posts: 13 Rating: (0)
Suggestion To thank Quote Answer

1/21/2011 2:45 PM	Rate (0)
Wüstenräuber Gold Expert Joined: 10/3/2008 Last visit: 9/8/2025 Posts: 996 Rating: (129)	"A couple of weeks ago" - then you should have the current version. Here you can see the downloads. Wüstenräuber

Suggestion To thank Quote Answer

1/21/2011 3:10 PM	Rate (0)
ToreK Posts: 13 Rating: (0)	This file contains the same FB versions I am using. 1.1 and 2.0. At the moment i need to call the ModbusCP FB 36 times to write/read all the information on the same port. This takes. as I said around 4 seconds. If the ModbusCP as you say is very fast. does this seem to be correct? (ModbusCP is at this test set in 20ms OB).
ToreK Posts: 13 Rating: (0)
Suggestion To thank Quote Answer

1/21/2011 3:46 PM	Rate (0)
Wüstenräuber Gold Expert Joined: 10/3/2008 Last visit: 9/8/2025 Posts: 996 Rating: (129)	Hi, you don't need so many FBs in your program like I described it above. The problem is, that the functions AG_LSEND/AG_LRECV can't be called unlimited. There is a maximal number of calls, which can be run parallel. This is described in the manual of the CPU/CP. I suppose this is the reason for the long time. You can also write a telegram traffic with wireshark. Then we can see what happens. Wüstenräuber
	Last edited by: Wuestenraeuber at: 1/21/2011 3:48 PM
Suggestion To thank Quote Answer

1/21/2011 4:02 PM	Rate (0)
ToreK Posts: 13 Rating: (0)	If I use BLKMOV and move the data from the "receive" DBs I would still need to change the unit 6 times(6 slavedIDs) and for each SlaveID I need to send 6 telegram to transmitt all data. This will call the ModbusCP block 36 times ( 36 messages) as far as I understand even by using the BLKMOV function?
ToreK Posts: 13 Rating: (0)
Suggestion To thank Quote Answer

1/25/2011 9:18 AM	Rate (0)
Wüstenräuber Gold Expert Joined: 10/3/2008 Last visit: 9/8/2025 Posts: 996 Rating: (129)	Yes, it's correct. In client mode 1 modbus telegram needs 2 CPU cycles for the handling (1 for sending and 1 for receiving). If the communication partner answers very slow, then it can takes a longer time. You could monitor it with wireshark. By the way, in server mode 1 modbus telegram needs 1 CPU cycle. The receiving and sending of the telegram is executed in 1 cycle. Wüstenräuber

Suggestion To thank Quote Answer

|

Share this page:

Share this page on...

Related topics by keywords:

ModbusTCP

Follow us on