Siemens PCS 7 Redundancy question

Hello to all,
I used to be a ControlLogix user but have started to look into Siemens PLCs. I found out from one of the online forums that PCS 7 is the closest to ControlLogix hence I have tried to concentrate my research in that direction. I have couple of issues that I have not been able to find in any manual:

- Which communication module on Siemens side offer similar functionality of CNBR modules from Rockwell? More importantly I am looking for the kind of media redundancy (Channel A and B) that CNBR offers. So that if one of my channel goes down down, the other one picks up. This is important to me because I do not want to put two communication modules in a rack.

- What makes the switchover happen in case of media fault? Is it hardware or software operated switchover?

Please help...

TP

I am working on my first S7-300 with redundancy now. I have redundant power supplies and processors with one I/O rack. The I/O rack contains a redundancy kit with two IM153-2. There is a profibus connection between CPU A and IM A and one between CPU B and IM B.

I purchased a redundancy bundle with softweare to set it up but have yet to get that far.

I have a manual PDF fully explaining it but it is too large to attach. Search for "software redundancy for Simatic S7-300 and S7-400" at http://support.automation.siemens.com/WW/llisapi.dll?func=cslib.csinfo&lang=en&objid=8859629&caller=view

Thanks Hassom,
Your answer is really helpful but I need to connect the CPU chassis with a "remote I/O" (I don't know how Siemens world calls it)chassis. And that is how I have used CNBR also. Two channels (A & B which are redundant to each other) of CNBR connects to another CNBR on a remote I/O chassis. I am sure you can connect them to PC or server but I have never used them like that.

I have one more thing to clear here.
- Does it mean that you have to have two CPs in one chassis in order to achieve media redundancy?

TP

I think you are talking about redundancy in media of the bus that connects the controllers to the remote I/O stations while I was talking about the bus that runs between controllers and PCs or servers.
In PCS7, they give definite name for each bus but I don't know the names exactly.

I will assume that you will use redundant controllers S7-400H and will not go for normal S7-400 (single controller) which can also work in PCS7.

If you will use S7-400H, you can put the 2 CPUs in a single rack (chassis).
It's a single rack (18 slots) but each CPU will take 9 isolated slots (not connected to the other 9 slots).

Each CPU will have a Profibus port on it which gives you an independent Profibus cable.
So, it means from one chassis you will have 2 Profibus cables.
Then you will have one of 2 choices,

1- To use a Y-Link which will take the 2 Profibus cables and produces a new single one which you can connect remote I/O stations to it.
The remote I/Os stations in this case could be of any kind but with a single Profibus interface because you have one cable.
You can also connect any DP-slave to this new Profibus cable.

2- To continue with the 2 Profibus cables (without Y-Link) and in this case the remote I/o stations must be ET200M with 2-Profibus interfaces (one for each cable).
Actually the 2 interfaces are separate modules linked together by an Active Bus Unit but these three elements are ordered as a single part under the name of RED BUNDLE.
This is the only way that achieves full redundancy over the I/O side.


There's something called Profinet which should replace Profibus in the future.
Profinet is similar to the Ethernet and can combine controllers, HMIs and I/O station in a single fast network.
There's ET200M which comes with 2 Profinet interfaces but as far as I know Profinet is not used with H systems.
You can check this with Siemens support.



SCADA or HMI bus ( I think it's called terminal bus in PCS7).
Similarly, you can use a single ETHERNET CP behind each CPU and you will have 2 independent ETHERNET cables out of the single rack (chassis).
You can also use 2 ETHERNET CPs behind each CPU and thus you will have 4 Ethernet cables out of the single rack (chassis).
There's a CP443-1 advanced which comes with 4 ports but I don't know if this single one would do the 2 CPs' job or not.

Siemens

I am also a controllogix user that has entered the Siemens world specifically PCS7.
The Siemens world offers a number of posibilities with regards to communications. The workhorse is Profibus DP. This bus is equivalent to controlnet although in my opinion much more reliable, greater support amongst suppliers, and is simple to setup. To achieve your redundancy goals, you have a number of options; multiple bus links, multiple CPUs, or Multiple communication processors. If you could provide more details, I can perhaps assist you. I think you stated that you are using PCS7. Therefore, you are most likely dealing with S7-400s that offer the most capability.
I just setup three pairs of S7-400H (redundant CPUs) with redundant profibus links and ET200s (remote I/O). After the sticker shock, I have found the system easy to configure and it works flawlessly. The nice thinh about Siemens is that although you pay top dollar, their stuff works really well. And their support is good and free as long as you maintain your system at relatively recent release status.
Again provide some design objectives and I could try to help.

I'm really curious,
I don't have much experience with the S7/PCS7 redundancy, but I'd like to know how how software redundancy really works. You say that redundancy switchover time is typical a few seconds and that Standby CPU is not executing same program which is the case with hardware redundany (hot standby). If I understood this, program is executed only in Mater CPU, but I don't know what is happening when Master fails, what ensures that Standby CPU is going to continue from same point when Master stopped?
Also I didn't quite understand the story with Y-link.
If you have a redundant configuration (2 CPUs), with Y link you connect two Profibus cables (redundant communication) into one that is connected to distribuited station ET200M. In this case this is not quite redundant solution since there is only one cable to remote I/Os.
Can you elaborate this more?
Thanks

Software redundancy (SWR).
It is used where brief unavailability (lasting a few processing cycles) of the system while switching from one station to another (from master to reserve station) can be tolerated by the process.

The solution is based on 2 CPUs A & B.
just to tell you, Software redundancy can help you to divide your plant or factory on 3 divisions (plants).
Plant A which is controlled by CPU A.
Plant B which is controlled by CPU B.
Plant C (high availability area) which is controlled by CPU A & CPU B (redundancy).
In such case, each CPU will have 2 different programs, non-redundant program and redundant program (Plant C program) which is typically the same on both CPUs.

Now since you can have such configuration, this means that there's no synchronization between both CPUs.
Even if you have a configuration of a single plant (high availability) which is controlled by both CPU A&B , there's still no synchronization between the 2 CPUs.
This means that at an instant of time, you may find CPU A executing OB1 while the CPU B has gone into OB35.

Now we come to the redundant link.
The redundant link is just a way of transferring data between both CPUs. In other words, the link is used to update the reserve CPU with latest data coming from master CPU.
For example, you may have all your setpoints in a data block and you will need both CPUs to have such setpoints.
The link will update reserve CPUs with setpoints.

Due to lack of synchronization and the rate of data transfer between both CPUs (MPI slowest ,Profibus faster, Ethernet fastest ) the reserve CPU (which receives setpoints from Master) will always have old setpoints (older values by no. of processing cycles).

You asked why the reserve CPU is not executing the program,
I don't know an exact answer but you must put into your consideration that,
1- Reserve CPU is always later than master with no. of seconds which means that there will be incorrect results of execution if both execute the program.
2- DP-slaves don't listen to reserve CPU and ignore their data which means that outputs are being taken from Master only.



Imagine now that you have totally lost Master CPU (loss of power),
Reserve CPU will take control in time that depends on the time of fault detection and the time of switching over of the DP-slaves but it will have older setpoints (That's the problem).
During the time of switching over, All outputs will be in freeze mode until the reserve CPU becomes the new master.
Now, the HMI needs to switch to the new master, this kind of switching is done using a VB or C script in WinCC.


Normally the time it takes to switch over from master to reserve depends on:

1- Amount of data you want to transfer through the redundant link.
2- Speed of redundant link (MPI,Prpfibus or Ethernet)
3- Number of DP slaves you have in your system.
4- Type of fault that causes switch over (loss of CPU or loss of DP slaves,.... .)



S7-400-H.
The redundant link or to be more precise the synchronization cables and synchronization modules gets both CPUs working in an impressive synchronization.
Both CPUs start together and end together. they load values together and get the results and outputs together but STILL the DP-slaves ignore the reserve CPU data and takes outputs from Master CPU.
In case of any fault, it's just a matter of few msecs and the reserve will take control.

Y-Link?
Yes, if use Y-Link, this doesn't mean you have achieved redundancy on Profibus or in I/Os side although you have achieved redundancy on CPUs side.
In many cases, the customer requires redundant CPUs (hot standby) and nothing more.
However there are many possible configurations of redundant I/Os in S7-400H.
If you are interested, please take a look at chapter 10 in S7-400H manual.
http://support.automation.siemens.c...view/en/1186523

Hello everyone!
I am new member.
I am trying to sovlve the redundant problem as below:
My system is redundant between 2 CPU315-2PN/DP via 2 CP342-5.
I have been configured redundance program.
When I turn the CPU1(StationA) off then all data is transfered to CPU2(StationB), I can control objects from wincc. But when I turn the CPU1 on, then turn the CPU2 off, the data is not transfer to cpu1 again. I have to restart the CPU1 then their status comeback.
That mean: CPU1 to CPU2: OK. But CPU2 to CPU1 not OK.
Any body have exprience to have me?
Many thank!

Hello everyone!
I am new member.
I am trying to sovlve the redundant problem as below:
My system is redundant between 2 CPU315-2PN/DP via 2 CP342-5.
I have been configured redundance program.
When I turn the CPU1(StationA) off then all data is transfered to CPU2(StationB), I can control objects from wincc. But when I turn the CPU1 on, then turn the CPU2 off, the data is not transfer to cpu1 again. I have to restart the CPU1 then their status comeback.
That mean: CPU1 to CPU2: OK. But CPU2 to CPU1 not OK.
Any body have exprience to have me?
Many thank!