(2)| 11/30/2023 6:16 AM | |
|
Joined: 10/7/2005 Last visit: 1/13/2026 Posts: 3054 Rating:
|
Generally speaking, the biggest bottleneck in all of this is the S7-300 backplane between CPU and CP. Unlike S7-400's, 300's do NOT have a dedicated K-bus (Communication bus) which now slows things down. The S7-Xplorer manual which you can download from www.iba-ag.com has benchmark figures in it for various CPU's and usage of onboard PN port vs external CP's. Although subject to total number of signals, you'll find that CP usage is about 4 times slower than using the CPU's PN port. Simplest fix is thus to use the CPU's PN port. If that is not possible - and presuming you use S7-Xplorer - you could give iba's S7-TCP/UDP Interface a go (which requires PLC programing though and a license for it). Other factors to consider particularly for S7-Xplorer are things like CPU loading (bear in mind the CPU's OS needs to assemble the iba requested signals) and how consecutive or fragmented the requested signals memory areas is (reading from plenty of different memory areas like I, Q, M, various DB's etc. does take longer). While this could be perhaps improved by pre-marshalling data into consecutive memory areas (e.g. all into one new DB), it requires PLC programming and PDA I/O Manager changes and you still have the backplane bottleneck. |
Last edited by: fritz at: 11/30/2023 06:17:20Cheers |
|
This contribution was helpful to
|
|
| 11/30/2023 7:14 AM | |
|
Joined: 10/18/2006 Last visit: 12/16/2025 Posts: 120 Rating:
|
Hi, just want to add, that the bandwidth of the S7-300 backplane bus (the "u-plugs") is limited to MPI speed, which are incredible 187,5 kBit/s. So even if you are connected via 100 MBit Ethernet or better, as long as you go over a CP and don't use a port directly on the CPU, you are pushed back to stone age. |
|
iba AG - Data Acquisition and Analysis the way it should be |
|
This contribution was helpful to
|
|
Follow us on