Krish

You might need more redundancy managers to achieve faster switch over of direction of ring

Kareto

Daga as I understand the functionality of Redudancy Manager it broke the ring and conect it only when ring breake in other place so there is no any ring in topology and X005s are fill them as in a bus. am I wrong?

daga

It is required to define the "ring" ports on all switches, which in my view is kind of automatic and default functionality. However, I will believe the manual when it says X005 cannot be used in a ring.

Yes I am pretty sure. I don't know which manuals are you referring to, but IRT is not a necessary requirement to be a redundancy manager. Please seeIKPI cataloguechapter on Profinet and Industrial ethernet (I had attaches the screenshot from that catalogue in earlier post) Scalance X200 manualhere. See table 3.1 section 3.1 page 25 for product features.

Hi Kareto

That is a safe decision to keep you hands clean in case of any trouble.

If it was something to be made very cheap thenoriginl ring would be ok.My reasoning is likeas follows. You are actually not using aplain ring. You are actually having a ring of stars. The PLC to HMIdata packets will not be coming on the ring except signals to other HMIs and other PLCs. They don´t get affected by any damage in the ring. Now coming to the ring part of the network , normally a single managed switch is sufficientto switch directions in plain ring network. But this switch will be only using the Rapid Tree Spanning system to switch over and switch ove r time will anywhere between a sec to 3 secs. It will be more towards a sec. If your network was something like office network with many trees and branches then this could go up to 3 secs but only with the some more managed. You can improve this time on by using a MRP enabled switch. In this case each switch has a MRP protocol which manages the media and immediately reacts to the change by broadcasting the change through the whole network. This will takeas low as 25ms.So if you have any application that needs to react at this speed then you have to use all managed switches with MRP enabled.

So in your case withnetwork you had posted, the HMIs and the PLCs connected to one switch would react almost immediatly with or without MRP.But HMIs reacting to PLC connected to different switches will react with a delay ofmax 3 secs if the ring network diconnects in between. With a MRP protocolthis delay will bearound 25ms (of course the HMIs will take there scanning time additionally).

Ethernet being universal you can test it all with very simple tools and sniffers whether this works the wayI mentioned. But the sales pitch of automation companies are very different and so it is sometimes difficult to sell the customer that the concept is not so complicated asit seems.Italso needs a thorough understanding of TCP/IP, Ethernet etc.For many contractually it is better to follow the Siemens books and recommendations. But if you are setting upjusta test system/or a similar arrangement , the simple oneshould be fine.



Krish